Recognition, Management, and Outcomes

of Depression in Primary Care

Gregory E. Simon, MD, MPH, Michael VonKorff, ScD

Objective: To evaluate the recognition, management,
and outcomes of depressed patients presenting in pri-
mary care.
Design: Epidemiologic survey with 12-month follow\x=req-\
up.
Setting: Primary care clinics of a staff-model health main-
tenance organization.
Patients and Main Outcome Measures: Consecu-
tive primary care attenders aged 18 to 65 years (n=1952)
were screened using the 12-item General Health Ques-
tionnaire (GHQ-12), and a stratified random sample
(n=373) completed a psychiatric assessment, including
the Composite International Diagnostic Interview (CIDI),
the 28-item GHQ, and a brief self-rated disability ques-
tionnaire (BDQ). Three-month follow-up assessment
(n=347) repeated the GHQ-28 and BDQ, and 12-month
follow-up (n=308) repeated the CIDI, GHQ-28, and BDQ.
Use of psychotropic drugs and mental health services was
assessed using computerized pharmacy and visit regis-
tration records.
Results: Structured interviews found 64 cases of cur-
rent major depression (weighted prevalence, 6.6%) and
58 cases of current subthreshold depression (weighted
prevalence, 8.8%). Of those with major depression, 64%
(n=41) were recognized as psychologically distressed by
the primary care physician, 56% (n=36) filled at least one
antidepressant prescription during the next 3 months,
and 39% (n=25) made at least one specialty mental health
visit. Compared with recognized cases, those with un-
recognized major depression were less symptomatic at
baseline (GHQ-28 score, 15.31 vs 11.07; P=.006) but
showed a similar rate of improvement over 12 months
(F test for difference in slopes, P=.93).
Conclusions: While many depressed primary care pa-
tients may go unrecognized and untreated, this group ap-
pears to have milder and more self-limited depression.
A narrow focus on increased recognition may not im-
prove overall outcomes. Treatment resources might be
best directed toward more intensive follow-up and re-
lapse prevention among those now treated.
(Arch Fam Med. 1995;4:99-105)

DEPRESSION
is a common1·2
disabling3·4 condition that
is most often managed in
primary care.5-6 Given the
personal suffering, lost
the first month.20"22 These apparent short¬
comings of current practice motivated the
Agency for Health Care Policy and Research
to identify depression treatment in primary
care as a priority area for development of

productivity,7 health care costs8·9 asso¬

From the Center for Health
Studies, Group Health
Cooperative of Puget Sound,
Seattle, Wash.
Downloaded from www.archfammed.com at CLOCKSS, on November 7, 2009
and
ciated with depression, appropriate treat¬
ment of depressed patients in primary care
should be a public health priority.
Judged by the standards of specialists,
current treatment of depression in primary
care appears to have significant shortcom¬
ings.10 Primary care physicians may rec¬
ognize only half of patients with major
depression1 µ17 and may initiate antidepres-
sant treatment in an even smaller portion
of cases.1819 Dosages of antidepressant
medication prescribed in primary care are
frequently lower than recommended.18"22
Many patients beginning antidepressant
treatment stop taking medication during
clinical practice guidelines.23
Available data, however, may not be
adequate to guide future clinical and edu-
caüonal interventions. Some of the data de¬
scribed above were collected a decade or
more ago1314·20 and may not accurately re¬
flect current practice. Most previous stud¬
ies contain no information on outcomes of
care,11131516·20·21 so the clinical impact of the
shortcomings described cannot be as-
 METHODS
STUDY SETTING
Group Health Cooperative of Puget Sound (GHC) is a staff-
model health maintenance organization serving an enrolled
population of approximately 390 000 in western Washington
State. GHC provides comprehensive health care (including pre¬
scribed medications) on a capitated basis. Specialty mental health
care is provided by the mental health service of GHC through
physician referral or self-referral. Over 95% of GHC physicians
providing primary care to adults are trained in family practice,
with most of the remainder trained in internal medicine. Over
90% are certified by appropriate specialty boards. Each full-
time physician provides outpatient and inpatient care for a de¬
fined panel of approximately 1600 patients. All subjects were
recruited from three GHC primary care clinics selected to rep¬
resent the socioeconomic range of Seattle area residents. The
GHC enrollment closely resembles the Seattle area population,
except for less representation of the high and low extremes of
income. Differences between the GHC enrollment and the US
population (there are fewer blacks and a higher educational
level in the GHC enrollment) primarily reflect the character¬
istics of Seattle area residents.
SCREENING
Consecutive primary care patients aged 18 to 65 years at¬
tending one of the study climes were asked to complete the
12-item General Health Questionnaire (GHQ-12), a brief
assessment of psychological distress.28 Criteria for ineligi-
bility included inability to speak English, acute medical ill¬
ness that precluded participation, or participation in the
study on a prior visit. Patients called from the waiting room
prior to completing screening were also excluded.
SECOND-STAGE ASSESSMENT
Screened patients were selected for further evaluation ac¬
cording to GHQ-12 scores: 10% of those scoring 2 or less,
35% of those scoring 3 or 4, and 100% of those scoring 5
or more. Interviews were conducted within 7 days of screen¬
ing either at the primary care clinic or in the patient's home.
The second-stage assessment included the following:
• The
Composite International Diagnostic Interview,29 a
structured diagnostic interview developed by the
sessed. The available data comparing outcomes among pri¬
mary care patients with recognized and unrecognized de¬
pression suggest that the unrecognized group is less
symptomatic and shows considerable short-term improve¬
ment without specific treatment.24"26 Current knowledge
leaves considerable uncertainty about the clinical signifi¬
cance of nonrecognition or failure to initiate antidepres¬
sant treatment.
See also pages 95 and 106
This report describes a longitudinal study of com¬
mon psychiatric disorders among primary care patients
of a large staff-model health maintenance organization.
Downloaded from www.archfammed.com at CLOCKSS, on November 7, 2009
World Health Organization that systematically evalu¬
ates diagnostic criteria of Diagnostic and Statistical
Manual of Mental Disorders, Fourth Edition (DSM-
IV), and international Statistical Classification of Dis¬
eases, 10th Revision (1CD-10).
• The Brief Disability Questionnaire, an 11-item self-
rated disability assessment adapted from the Medical Out¬
comes Study disability questionnaire short form.30
• The 28-Item General Health Questionnaire (GHQ-28),28
a self-report assessment of psychiatric symptoms fre¬
quently seen in primary care.
For each subject selected, the primary care physi¬
cian completed a brief rating of medical comorbidity, phy¬
sician's assessment of psychological disorder, and treat¬
ment prescribed at the index visit.
3-MONTH FOLLOW-UP
Subjects completing the initial interview were contacted
by telephone 3 months later for repeat administration of
the GHQ-28, Brief Disability Questionnaire, and a brief
questionnaire concerning health care use. Use of psycho-
tropic medications during the 3 months following treat¬
ment period (including prescribed dosage and duration
of antidepressant treatment) was examined using GHCs
computerized pharmacy records.22 Prior internal surveys
have shown that GHC enrollees fill over 95% of outpa¬
tient prescriptions (including antidepressant prescrip¬
tions) at GHC pharmacies (data available from the au¬
thors on request). Use of specialty mental health services
was assessed using both computerized visit registration
records (for services within the health maintenance orga¬
nization) and patient questionnaires (for services outside
the health maintenance organization).
12-MONTH FOLLOW-UP
Subjects completing the initial interview were contacted
for an in-person assessment 12 months later, including re¬
peat administration of the diagnostic interview, GHQ-28,
and Brief Disability Questionnaire.
DATA ANALYSIS
All analyses were performed using SPSS software (SPSS Ine,
Chicago, 111). Prevalence estimates incorporate sampling
weights,31 but other results are based on unweighted data.
The study included baseline diagnostic assessment and
follow-up examination of treatment received, clinical out¬
comes, and functional status. Data reported here are from
the Seattle, Wash, site of the international collaborative
survey of Psychological Problems in General Health Care,
organized by the Division of Mental Health of the World
Health Organization, Geneva, Switzerland.27
RESULTS
STUDY SAMPLE
Of the 2592 patients approached for screening, 2110 were
eligible and 1962 (93% of those eligible) completed the
GHQ-12. Of the 615 subjects identified as eligible for the
second-stage assessment, 373 (61%) completed the in¬
terview. Lack of time was the primary reason given for
refusal. Demographic characteristics of those screened
and those interviewed are shown in Table 1. The high
proportion of female subjects reflects primarily the greater
use of primary care services among women in the eli¬

gible age range. The high educational level reflects the

characteristics of the GHC enrollment. Table 2 com¬
pares the 373 subjects who completed the second-stage
interview with the 242 eligible subjects who refused or
failed to show up for interview appointments. The GHCs
computerized pharmacy and registration systems were
used to assess the portion in each group that visited GHCs
Mental Health Service or received a psychotropic pre¬
scription during the 3 months following screening. Com¬
pleters were slightly older, but the two groups did not
differ significantly in gender, GHQ-12 score, or likeli¬
hood of receiving subsequent mental health care or a psy¬
chotropic drug prescription.
Follow-up assessments were completed on 347 pa¬
tients (93.0%) at 3 months and 305 patients (81.8%) at
12 months. Likelihood of completing follow-up was
slightly lower among the youngest patients but was not *Data on marital status, employment, and educational attainment were
related to sex, presence of depression diagnosis at base¬ not collected for those who were screened but not interviewed.
line, baseline GHQ-28 score, or baseline score on the Brief
Disability Questionnaire.
PREVALENCE

Based on diagnostic interviews, 64 subjects satisfied

DSM-IV criteria for current major depressive episode,
yielding a weighted prevalence of 6.6%. Subthreshold de¬
pression was defined as the presence of either depressed
mood or dysthymia plus two or three additional posi¬
tive symptoms from the nine DSM-IV criteria. By this stan¬
dard, an additional 58 subjects had current subthresh-
old depression, for a weighted prevalence of 8.8%. Because
the stratified random sampling procedure oversampled
those with higher GHQ-12 scores, the actual number of
cases in the interviewed sample was not directly propor¬
tional to weighted prevalence in the screened sample. Pa¬
tients with current major depression and those with sub-
*
Based on t test for age and GHQ-12 score, 2 analysis for all other
variables.
threshold depression were similar in mean age (38.2 and tGHQ-12 indicates 12-item General Health Questionnaire.
37.2 years, respectively) and gender distribution (73% ¿.During the 3 months following screening.
[n=47] and 77% [n=45] female, respectively). The ma¬
jor depression group had higher mean scores than the
subthreshold depression group on the baseline GHQ-28 tients with major or minor depression, likelihood of rec¬
(13.7 vs 10.0, F=11.65, P-C001) and the baseline Brief ognition was not significantly related to patient gender
Disability Questionnaire (9.2 vs 7.2, F=3.6, P=.05). Over¬ or education but was slightly lower among younger pa¬
all, 15.4% of consecutive primary care patients met cri¬ tients. The portion of patients with major depression or
teria for any current depressive disorder (major depres¬ subthreshold depression recognized as cases was 28%
sive episode or subthreshold depression). (eight of 29) among those aged 18 to 30 years, 48% (15
of 31) among those aged 31 to 40 years, 57% (29 of 51)
RECOGNITION among those aged 41 to 50 years, and 55% (six of 11)
among those aged 51 to 65 years (Mantel-Haenszel 2 test
Most patients with major depression were recognized by for linear association, 5.36; df=\; P=.02). Recognition was
their primary care physicians. Based on questionnaires related to both severity of psychiatric symptoms and dis¬
completed at the index visit, primary care physicians rec¬ ability. Compared with patients who had unrecognized
ognized 41 (64%) of 64 patients with major depression depression, patients with any recognized depressive dis¬
and 17 (29%) of 58 patients with subthreshold depres¬ order had higher mean scores on the baseline GHQ-28
sion as "cases'' of psychological disorder. Among pa- (12.79 vs 10.08; F=6.15; dj=l,120; P=.01) and Brief Dis-

Downloaded from www.archfammed.com at CLOCKSS, on November 7, 2009

 Recognized (n=34)
Unrecognized (n=16)
0 0
Baseline 3 mo 12
Figure 1. Course of scores on the 28-item General Health Questionnaire
(GHQ-28) for patients with major depression at baseline according to
physician recognition.
ability Questionnaire (8.95 vs 7.06; F=3.41; d/=l,118;
P=.07).
TREATMENT
Of patients with major depressive disorder, 25 (39%) re¬
ported use of antidepressant medications at the time of
the diagnostic interview, and computerized pharmacy
records showed that a total of 36 (56%) received anti¬
depressant medications at some time during the 3 months
following screening. Likelihood of antidepressant use was
not significantly related to patient gender or education
level but was higher among older patients (31% [four of
13] among those aged 18 to 30 years, 57% [eight of 14]
among those aged 31 to 40 years, 64% [20 of 31] among
those aged 41 to 50 years, and 67% [four of six] among
those aged 50 to 65 years [Maentel-Haenszel 2 test for
linear association, 3.76; df=l; P=.05]). Among subjects
with major depression at baseline who completed the
3-month follow-up, 23 (39%) of 59 reported at least one
specialty mental health visit since screening. Using ei¬
ther indicator of treatment (use of antidepressant medi¬
cation or use of specialty mental health services), 39 (66%)
of 59 patients with major depression at baseline re¬
ceived some treatment during the next 3 months.
Likelihood of receiving treatment was strongly re¬
lated to severity of illness. Compared with untreated pa¬
tients, patients with major depression receiving antide¬
pressant medication were more symptomatic (baseline
GHQ-28 score, 15.31 vs 11.07; F=7.98;d/= 1,62; P=.006)
and more disabled (baseline Brief Disability Question¬
naire score, 10.54 vs 7.46; F=4.51; d/=l,61; P=.04). Com¬
pared with those using no specialty mental health ser¬
vices, patients with major depression using any specialty
mental health services were also more symptomatic (base¬
line GHQ-28 score, 16.00 vs 11.97; F=5.86; d/=l,57;
P=.02) and more disabled (baseline Brief Disability Ques¬
tionnaire score, 10.96 vs 7.80; F=4.54; d/=l,56; P=.04).
Consistent with these findings, treatment rates were lower
for patients with subthreshold depression. Of 58 pa¬
tients with subthreshold depression at baseline, 11 (19%)
received antidepressant medication during the 3 months
after screening, and seven (13%) of 52 who completed
Downloaded from www.archfammed.com at CLOCKSS, on November 7, 2009
Treated ( =28)
Untreated (n=22)
mo Baseline 3 mo 12 mo
Figure 2. Course of scores on the 28-item General Health Questionnaire
(GHQ-28) for patients with major depression at baseline according to
receipt of antidepressant treatment.
the 3-month follow-up reported use of specialty mental
health services during the 3 months after screening.
Most patients treated with antidepressant medica¬
tions received therapy consistent with the recommen¬
dations for depression treatment in primary care re¬
cently published by the Agency for Health Care Policy
and Research.23 Among the 36 patients with major de¬
pression who received antidepressant medications, 28
(78%) received dosages within the recommended ranges
(eg, at least 75 mg/d of imipramine), and 27 (75%) con¬
tinued to refill antidepressant prescriptions for at least
90 days. Both dosage and duration of treatment met these
standards in 22 (61%) of 36 cases.
OUTCOMES
Patients with both major and minor depression showed
considerable improvement in depression and disability
during the follow-up period. Among patients with ma¬
jor depression at baseline, 50 completed both the 3- and
12-month follow-up assessments. In that sample, the mean
GHQ-28 score fell from 14.12 at baseline to 6.04 at 3
months to 5.81 at 12 months. At the 12-month assess¬
ment, 15 patients (30%) continued to satisfy criteria for
major depressive disorder, four patients (8%) met crite¬
ria for minor depression, and 31 patients (62%) did not
satisfy criteria for any depressive disorder. Among pa¬
tients with minor depression at baseline, the mean
GHQ-28 score fell from 8.66 at baseline to 4.78 at 3
months to 3.90 at 12 months.
Patients with unrecognized and/or untreated de¬
pression showed rates of improvement similar to those
of patients with recognized and/or treated depression.
Figure I displays the symptomatic course of patients
with major depression at baseline according to recogni¬
tion by the primary care physician. Patients with recog¬
nized and unrecognized depression showed similar rates
of improvement during the first 3 months of follow-up,
and both groups remained relatively stable thereafter. Mul¬
tivariate analysis of variance indicated a significant ef¬
fect of time (F=29.78; d/=2,44; P<.001) but no signifi¬
cant groupX time interaction (F=.07; d/=2,44; P=.93). The
two groups also showed similar likelihoods of remis-
sion by the 12-month assessment. The likelihood of com¬
plete remission (no depressive diagnosis at 12-month
follow-up) was 60% (21 of 35) for the recognized group
and 68% (10 of 15) for the unrecognized group ( 2=1.55,
d/=l, P=.45). Figure 2 displays the symptomatic course
of patients with major depression at baseline classified
according to treatment with antidepressant medications
during the 3 months following screening. Comparison
using multivariate analysis of variance again found no
evidence for difference in improvement over time (F=0.58;
d/=2,44;P=.56).
COMMENT
These data allow a more comprehensive view of depres¬
sion treatment in primary care than do most earlier re¬
ports. The numbers of patients with major (n=64) and
minor (n=58) depression allow accurate estimates of rec¬
ognition rates, treatment rates, and severity of illness. Data
on use of mental health services and psycho tropic drugs
give a more detailed view of treatment over time. Fol¬
low-up assessments allow examination of the clinical out¬
comes of underrecognition and undertreatment.
Our findings confirm the importance of depression
in primary care but give a somewhat more optimistic view
of treatment than previously described.10 Consistent with
most previous reports, 6.6% of consecutive primary care
patients satisfied formal criteria for major depressive dis¬
order, and a slightly larger number had significant de¬
pressive symptoms that fell below the DSM-IV diagnos¬
tic threshold. Of patients with major depression,
approximately 60% were recognized as psychologically
ill by the treating primary care physician, and nearly two
thirds received either antidepressant medication or some
specialty mental health treatment over the following 3
months. Recognition and initiation of treatment were both
associated with severity of illness. The depressed pri¬
mary care patients in this sample showed relatively good
short-term outcomes, and rates of improvement were not
related to either physician recognition or receipt of an¬
tidepressant treatment.
While reviews often describe lower rates of recog¬
nition of depression in primary care, the recognition rate
in this sample is consistent with rates in most previous
studies using formal diagnostic criteria to identify ma¬
jor depression. VonKorff et al,11 Gerber et al,12 Wells et
al,16 and Ormel et al23 all reported recognition rates of
50% to 65% for primary care patients with depressive dis¬
orders diagnosed by structured interview. In each of these
studies, recognition rates were based on specific ques¬
tions to treating physicians regarding psychological dis¬
orders. Substantially lower recognition rates reported in
other studies probably reflect significant méthodologie
differences. The 26% recognition of major depression re¬
ported by Schulberg et al14 was based on chart notations
of psychiatric diagnoses. Direct questioning of treating
physicians is likely to yield higher estimates of recogni¬
tion. In this sample, the agreement between recognition
as measured by physician questionnaire and the portion
of patients receiving treatment supports the question¬
naire method as an estimate of clinically meaningful rec¬
ognition. In studies by Nielsen and Williams,13 Magruder-
Downloaded from www.archfammed.com at CLOCKSS, on November 7, 2009
Habib et al,19 and Zung et al32 that reported recognition
rates of 15% to 43%, patients were identified by depres¬
sion screening scales without formal diagnostic assess¬
ment. The higher prevalence rates estimated by studies
using the symptom scale method (13% to 17%) prob¬
ably reflect a lower threshold for diagnosis of depres¬
sion. As seen in this sample, inclusion of less severe
depression significantly reduces apparent recognition
rates. Given that treatment efficacy is only established
for patients above the major depression threshold, rec¬
ognition rates based on diagnosed major depression seem
more clinically relevant.
The relationship observed in this sample between
the recognition of depression and short-term clinical out¬
comes is also similar to that seen in earlier studies. In a
sample of Dutch primary care patients reported by Ormel
et al,25 physician recognition of depression was not re¬
lated to episode duration or severity of illness at 14-
month follow-up.17 Schulberg and colleagues24 found no
relationship between recognition and 6-month out¬
comes. According to available data,17'24 26 primary care pa¬
tients with unrecognized depression are less severely ill
than those whose depression is recognized, and they have
similar or superior short-term outcomes. Data from the
physicians are more likely to diagnose
and treat those with more severe
illness and greater disability
Medical Outcomes Study18 show that the likelihood of
antidepressant treatment also increases with severity of
illness. Observational studies of patients drawn from psy¬
chiatric practice33·34 find a similar pattern: untreated de¬
pressive episodes show a more benign course.
We do not interpret these findings as evidence against
the efficacy of depression treatment in primary care. In¬
stead, these findings may reflect treatment matching by
the primary care physician. Among the group of pri¬
mary care patients presenting with depressive illness, phy¬
sicians are more likely to diagnose and treat those with
more severe illness and greater disability. Favorable short-
term outcomes among unrecognized and untreated pa¬
tients suggest that primary care physicians are reason¬
ably accurate in identifying depressed patients who have
a high likelihood of spontaneous recovery.
Our data on intensity of antidepressant treatment
do appear inconsistent with data of earlier reports. Pre¬
vious descriptions of antidepressant treatment in pri¬
mary care,18 including two studies from the same health
20
maintenance organization,21·22 found that the majority of
patients receive subtherapeutic antidepressant dosages
or stop treatment during the first month. Different re¬
sults in this sample may reflect the small number of pa¬
tients treated with antidepressant medications (n=36).
Alternatively, the higher rates of adequate antidepres¬
sant treatment seen here may reflect differences in sample
selection. Previous studies19"22 have typically examined
patients beginning courses of antidepressant treatment,
including a large portion with transient depression. In
contrast, a prevalence-based sample typically includes a
larger portion of chronic cases. Consequently, the cross-
sectional method used in this study probably selected a
much higher portion of persistent cases than included
in earlier incident samples. If intensity of treatment is re¬
lated to persistence of depression, a cross-sectional sample
the greatest burden of preventable
depression morbidity may lie in
subsequent relapse or recurrence
would also yield higher estimates of treatment inten¬
sity. Consistent with this explanation, the cross-
sectional Medical Outcomes Study18 found a lower rate
of subtherapeutic dosing (39%) than usually seen in in¬
cident samples.
These optimistic findings may raise questions
about generalizability to other primary care settings.
This sample was drawn from one health care system
serving a population that was largely white, employed,
and well educated. Other ethnic or socioeconomic
groups may experience considerably poorer outcomes.
Most participating primary care providers were trained
in family medicine and may have received more inten¬
sive mental health training than physicians in other
specialties. The high rate of specialty board certifica¬
tion among study physicians may indicate a higher
level of mental health training or knowledge than gen¬
erally found among primary care physicians. Previous
studies, however, have suggested lower rates of recog¬
nition and poorer outcomes for depressed patients
treated in prepaid health care systems.16·35
These findings do suggest that improving the out¬
comes of depressed primary care patients may require a
shift in focus. By the traditional measures of physician
recognition and short-term outcomes, the gap between
knowledge and current practice may not be as wide as
often described.10 Most cases of major depression were
recognized by treating physicians and received some level
of specific treatment. Physician recognition, likelihood
of initiating treatment, and intensity of treatment were
all strongly related to severity of illness. Most patients
(including those unrecognized and untreated) had rela¬
tively good short-term outcomes. These results, how¬
ever, reflect short-term outcomes of patients attending
primary care appointments and may give an incomplete
picture of overall outcomes.
First, good short-term outcomes may not reflect long-
term prognosis and risk of relapse or recurrence. In this
sample, over 75% of patients with major depression re¬
ported at least one previous depressive episode. Consis¬
tent with this, longitudinal studies of epidemiologie
samples36 and treated samples37·38 find high short-term
remission rates but high risk of subsequent relapse or re¬
currence. Consequently, the greatest burden of prevent¬
able depression morbidity may lie in subsequent relapse
or recurrence, not in persistence of the index episode. A
narrow focus on improving short-term remission rates
may have minimal impact on overall morbidity.
Second, this cross-sectional sample may not
include patients who initiated treatment for depres¬
sion but did not return for follow-up. Our previous
study of primary care patients initiating antidepressant
Downloaded from www.archfammed.com at CLOCKSS, on November 7, 2009
treatment22 found that a substantial minority of
patients who stop antidepressant treatment during the
first month make no follow-up visits during the next 6
months. Without clinical follow-up data, we cannot
determine whether this failure to return indicates
spontaneous recovery or discouragement and persis¬
tent depression. This group of patients will not be
reached by efforts to increase physician recognition or
to boost intensity of antidepressant treatment. Improv¬
ing outcomes in patients currently receiving no
follow-up care willrequire systematic and active
monitoring after treatment initiation.
Our results and those of earlier primary care
samples argue for a more focused approach toward
improving depression treatment. Patients with unrec¬
ognized depression were less severely ill than those
with recognized depression, and they remained less ill
at follow-up. Consequently, efforts to improve detec¬
tion through physician education or routine screening
may not improve overall outcomes. Similarly, patients
with major depression not treated with antidepressant
medication were less ill at baseline and improved con¬
siderably over 3 months. Increasing the portion of
patients treated with antidepressant medications might
only result in treatment of less severely ill patients, in
whom the efficacy of antidepressants is not well estab¬
lished.39·40 Good short-term outcomes in the entire
sample suggest that many depressed primary care
patients may not need more intensive short-term treat¬
ment. Resources might be more profitably directed at
ensuring adequate follow-up care for patients cur¬
rently beginning treatment and at reducing the risk of
subsequent relapse or recurrence.
Accepted for publication November 8, 1994.
This study was supported by grant MH47765 from
the National Institute of Mental Health, Bethesda, Md.
These data were collected as part of the Psychologi¬
cal Problems in General Health Care Study, coordinated
by the Division of Mental Health of the World Health Or¬
ganization, Geneva, Switzerland. Participating investiga¬
tors include Orhan Ozturk, MD, and Murat Rezaki, MD,
Ankara, Turkey; Costas Stefanis, MD, and Venetsanos
Mavreas, MD, PhD, Athens, Greece; S. M. Channabasa-
vanna, MBBS, MD, and T. G. Sriram, MBBS, MD, Ban¬
galore, India; Hanfried Helmchen, MD, and Michael Lin¬
den, PD, Berlin, Germany; Wim van den Brink, MD, and
Bea Tiemens, PhD, Groningen, the Netherlands; Micheal
Olawatura, MD, and Oye Gureje, MD, Ibadan, Nigeria;
Otto Benkert, MD, and Wolfgang Maier, MSc, MD, Mainz,
Germany; Richard Gater, MD, and Steven Kisely, MB, ChB,
Manchester, England; Yoshibume Nakane, MD, and Shu-
nichirou Michitsuji, MD, Nagasaki, Japan; Yves Lecru-
bier, MD, and Patrice Boy er, MD, Paris, France; Jorges-
Alberto Costa e Silva, MD, and Louis Villano, MD, Rio de
Janeiro, Brazil; Ramon Florenzano, MD, MPH, and Julia
Acuna, MD, Santiago, Chile; Gregory Simon, MD, MPH,
and Michael VonKorff, ScD, Seattle, Wash; Yan He-Quin,
MD, and Xaio Shi Fu, MD, Shanghai, China; and Michelle
Tansella, MD, and Cesario Bellantuono, MD, Verona, Italy.
The study advisory group includes Jorges-Alberto Costa e
Silva, MD, David P. Goldberg, MA, BM, Yves Lecrubier,
MD, Johan Ormel, PhD, Michael VonKorff, ScD, and Hans-
Ulrich Wittchen, PhD. Coordinating staffai World Health
Organization headquarters include Norman Sartorius, MD,
PhD, and T. Bedirhan Ustun, MD, PD.
Chester Pabiniak assisted with computer program¬
ming.
Reprint requests to Center for Health Studies, #1600,
1730 Minor Ave, Seattle, WA 98101-1448 (Dr Simon).
REFERENCES
1. Regier DA, Boyd JH, Burke JD, et al. One-month prevalence of mental disor-
ders in the United States. Arch Gen Psychiatry. 1988;45:977-986.
2. Kessler RC, McGonagle KA, Zhao S, et al. Lifetime and 12-month prevalence
of DSM-III-R psychiatric disorders in the United States: results from the Na-
tional Comorbidity Survey. Arch Gen Psychiatry. 1994;51:8-19.
3. Wells K, Stewart A, Hays R, et al. The functioning and well-being of depressed
patients: results from the Medical Outcomes Study. JAMA. 1989;262:914\x=req-\
919.
4. VonKorff M, Ormel J, Katon WJ, Lin EHB. Disability and depression among
high utilizers of health care. Arch Gen Psychiatry. 1992;49:91-100.
5. Schurman RA, Kramer PD, Mitchell JB. The hidden mental health network:
treatment of mental illness by nonpsychiatrist physicians. Arch Gen Psychia-
try. 1985;42:89-94.
6. Regier D, Narrow WE, Rae DS, Mandersheid RW, Locke BZ, Goodwin FK. The
de facto US mental and addictive disorders service system: epidemiologic catch-
ment area prospective 1-year prevalence rates of disorders and services. Arch
Gen Psychiatry. 1993;50:85-94.
7. Broadhead WE, Blazer DG, George LK, Tse CK. Depression, disability days and
days lost from work in a prospective epidemiologic survey. JAMA. 1990;264:
2524-2528.
8. Simon G. Psychiatric disorder and functional somatic symptoms as predictors
of health care use. Psychiatr Med. 1992;10:49-60.
9. Greenberg P, Stiglin LE, Finkelstein SN, Berndt ER. The economic burden of
depression in 1990. J Clin Psychiatry. 1993;54:405-418.
10. Eisenberg L. Treating depression and anxiety in primary care: closing the gap
between knowledge and practice. N Engl J Med. 1992;326:1080-1084.
11. VonKorff M, Shapiro S, Burke JD, et al. Anxiety and depression in a primary care
clinic: comparison of Diagnostic Interview Schedule, General Health Question-
naire, and practitioner assessments. Arch Gen Psychiatry. 1987;44:152-156.
12. Gerber PD, Barrett J, Barrett J, Manheimer E, Whiting R, Smith R. Recognition
of depression by internists in primary care: a comparison of internist and 'gold
standard' psychiatric assessments. J Gen Intern Med. 1989;4:7-13.
13. Nielsen AC, Williams TA. Depression in ambulatory medical patients: preva-
lence by self-report questionnaire and recognition by nonpsychiatric physi-
cians. Arch Gen Psychiatry. 1980;37:999-1004.
14. Schulberg HC, Saul M, McClelland M, Ganguli M, Christy W, Frank R. Assess-
ing depression in primary medical and psychiatric practices. Arch Gen Psy-
chiatry. 1985;42:1164-1170.
15. Jones L, Badger LW, Ficken RP, Leeper JD, Anderson RL. Inside the hidden
mental health network: examining mental health care delivery of primary care
physicians. Gen Hosp Psychiatry. 1987;9:287-293.
16. Wells K, Hays RD, Burnam MA, Rogers W, Greenfield S, Ware JE. Detection of
depressive disorder for patients receiving prepaid or fee-for-service care: re-
sults from the Medical Outcomes Study. JAMA. 1989;262:3298-3302.
17. Ormel J, van den Brink W, Koeter MW, van der Meer D, van de Willige G,
Wilmink FW. Recognition, management, and outcome of psychological disor-
ders in primary care: a naturalistic follow-up study. Psychol Med. 1990;20:
909-923.
Downloaded from www.archfammed.com at CLOCKSS, on November 7, 2009
18. Wells K, Katon W, Rogers B, Camp P. Use of minor tranquilizers and antide-
pressant medications by depressed outpatients: results from the Medical Out-
comes Study. Am J Psychiatry. 1994;151:694-700.
19. Magruder-Habib K, Zung WWK, Feussner JR, Ailing WC, Saunders WB, Stevens
HA. Management of general medical patients with symptoms of depression.
Gen Hosp Psychiatry. 1989;11:201-206.
20. Johnson DAW. A study of the use of antidepressant medication in general prac-
tice. Br J Psychiatry. 1974;125:186-192.
21. Katon W, VonKorff M, Lin E, Bush T, Ormel J. Adequacy and duration of an-
tidepressant treatment in primary care. Med Care. 1992;30:67-76.
22. Simon G, VonKorff M, Wagner EH, Barlow W. Patterns of antidepressant use
in community practice. Gen Hosp Psychiatry. 1993;15:399-408.
23. Depression Guideline Panel. Depression in Primary Care: Clinical Practice Guide-
line. Rockville, Md: Agency for Health Care Policy and Research, US Dept of
Health and Human Services; 1993. Publications AHCPR 93-0550 and 93\x=req-\
0551.
24. Schulberg HC, McClelland M, Gooding W. Six-month outcomes for medical
patients with major depressive disorders. J Gen Intern Med. 1987;2:312-317.
25. Ormel J, Koeter MW, van den Brink W, van de Willige G. Recognition, man-
agement, and course of anxiety and depression in general practice. Arch Gen
Psychiatry. 1991;48:700-706.
26. Coyne J, Schwenk TL, Fechner-Bates S. Non-detection of depression by pri-
mary care physicians reconsidered. Gen Hosp Psychiatry. In press.
27. Sartorius N, \l=U"\st\l=u"\nTB, Costa e Silva J-A, et al. An international study of psy-
chological problems in primary care: preliminary report from the World Health
Organization Collaborative Project on 'Psychological Problems in General Health
Care.' Arch Gen Psychiatry. 1993;50:819-824.
28. Goldberg D. Manual of the General Health Questionnaire. Windsor, England:
National Foundation for Educational Research; 1978.
29. Robins LN, Wing J, Wittchen H-U, et al. The Composite International Diag-
nostic Interview: an epidemiologic instrument suitable for use in conjunction
with different diagnostic systems in different cultures. Arch Gen Psychiatry.
1988;45:1069-1077.
30. Stewart A, Hays R, Ware J. The MOS Short-form General Health Survey: re-
liability and validity in a patient population. Med Care. 1988;26:724-735.
31. Cochran W. Sampling Techniques. New York, NY: John Wiley & Sons Inc;
1977.
32. Zung W, Magill M, More JT, George DT. Recognition and treatment of depres-
sion in a family medical practice. J Clin Psychiatry. 1983;44:3-6.
33. Brugha T, Bebbington PE, MacCarthy B, Sturt E, Wykes T. Antidepressants
may not assist recovery in practice: a naturalistic prospective survey. Acta Psy-
chiatr Scand. 1992;86:5-11.
34. Coryell W, Akiskal HS, Leon AC, et al. The time course of nonchronic major
depressive disorder: uniformity across episodes and samples. Arch Gen Psy-
chiatry. 1994;51:405-410.
35. Wells K, Burnam M, Rogers W. Course of depression for adult outpatients:
results from the Medical Outcomes Study. Arch Gen Psychiatry. 1992;49:788\x=req-\
794.
36. Angst J. Epidemiology of depression. Psychopharmacology. 1992;106(suppl):
71-74.
37. Shea M, Elkin I, Imber SD, et al. Course of depressive symptoms over follow-
up: findings from the National Institute of Mental Health Treatment of Depres-
sion Collaborative Research Program. Arch Gen Psychiatry. 1992;49:782-787.
38. Keller M, Lavori P, Mueller T, et al. Time to recovery, chronicity, and levels of
psychopathology in major depression: a 5-year prospective follow-up of 431
subjects. Arch Gen Psychiatry. 1992;49:809-816.
39. Elkin I, Shea T, Watkins JT, et al. National Institute of Mental Health Treatment
of Depression Collaborative Research Program. Arch Gen Psychiatry. 1989;
46:971-982.
40. Paykel ES, Hollyman JA, Freeling P, Sedgwick P. Predictors of therapeutic ben-
efit from amitriptyline in mild depression: a general practice placebo-
controlled trial. J Affect Disord. 1988:14:83-95.