Customize your JAMA Network experience by selecting one or more topics from the list below.
Das AK, Olfson M, Gameroff MJ, et al. Screening for Bipolar Disorder in a Primary Care Practice. JAMA. 2005;293(8):956–963. doi:10.1001/jama.293.8.956
Context Bipolar disorder consists of episodes of manic and depressive symptoms.
Efforts to screen for depression in a primary care setting without assessment
of past manic symptoms can lead to incorrect diagnosis and treatment of bipolar
Objectives To screen for bipolar disorder in adult primary care patients and to
examine its clinical presentation and effect on functioning.
Design, Setting, and Participants A systematic sample of 1157 patients between 18 and 70 years of age
who were seeking primary care at an urban general medicine clinic serving
a low-income population. The study was conducted between December 2001 and
Main Outcome Measures Prevalence of bipolar disorder, its treatment and patient functioning.
Study measures included the Mood Disorder Questionnaire, the PRIME-MD Patient
Health Questionnaire, the Medical Outcomes Study 12-Item Short Form health
survey, the Sheehan Disability Scale, data on past mental health treatments,
and a review of medical records and International Classification
of Diseases, Ninth Revision codes for each visit dating from 6 months
prior to the screening day.
Results The prevalence of receiving positive screening results for lifetime
bipolar disorder was 9.8% (n = 112; 95% confidence interval, 8.0%-11.5%)
and did not differ significantly by age, sex, or race/ethnicity. Eighty-one
patients (72.3%) who screened positive for bipolar disorder sought professional
help for their symptoms, but only 9 (8.4%) reported receiving a diagnosis
of bipolar disorder. Seventy-five patients (68.2%) who screened positive for
bipolar disorder had a current major depressive episode or an anxiety or substance
use disorder. Of 112 patients, only 7 (6.5%) reported taking a mood-stabilizing
agent in the past month. Primary care physicians recorded evidence of current
depression in 47 patients (49.0%) who screened positive for bipolar disorder,
but did not record a bipolar disorder diagnosis either in administrative billing
or the medical record of any of these patients. Patients who screened positive
for bipolar disorder reported worse health-related quality of life as well
as increased social and family life impairment compared with those who screened
Conclusions In an urban general medicine clinic, a positive screen for bipolar disorder
appears to be common, clinically significant, and underrecognized. Because
of the risks associated with treating bipolar disorder with antidepressant
monotherapy, efforts are needed to educate primary care physicians about the
screening, management, and pharmacotherapy of bipolar disorders.
There have been numerous reports on the prevalence1-15 and
treatment16,17 of major depression
in primary care. Routine screening for depression is now recommended in primary
care settings that have the capacity to provide effective management of these
cases.18 In spite of this progress, little
attention has been given to primary care patients who have current depression
and past episodes of hypomania or mania, history that may indicate bipolar
disorder and a need for specialized treatment.
Bipolar disorders comprise a clinical spectrum including bipolar I disorder,
bipolar II disorder, and cyclothymia.19 One
or more episodes of manic symptoms such as euphoric or irritable mood, racing
thoughts, a decreased need for sleep, talkativeness, and excessive involvement
in risk-taking activities are necessary to diagnose a bipolar disorder. Hypomania
and mania share the same symptom criteria but differ in episode duration and
impairment severity.19 Hypomania may be more
prevalent than mania, but also more difficult to detect through screening.20
Patients with bipolar disorders are more likely to present during an
episode of depression than hypomania or mania.21,22 In
one suburban family practice, 28 of 108 consecutive adult patients (25.9%)
with depressive or anxiety disorders assessed by a physician had a lifetime
history of hypomania or mania.23 Seventy-seven
(71.4%) had not previously been diagnosed or treated for bipolar disorders.23 Because of the 2-stage design of this study, it was
not possible to estimate the prevalence of bipolar disorder among participants.
A history of hypomania or mania can easily be missed in this setting since
primary care physicians may not routinely ask patients about past mood elevation.24
Determining whether depression is part of a depressive or bipolar disorder
is essential for appropriate pharmacological management. Treatment of patients
with bipolar disorder with unopposed antidepressants, such as serotonin reuptake
inhibitors, risks precipitating mania, hypomania, mixed affective states,
and rapid cycling between depression and mania.25,26 This
risk is reduced with a mood-stabilizing agent such as lithium or divalproex
or an antipsychotic medication. The American Psychiatric Association practice
guideline cautions against antidepressant monotherapy in the management of
The specific aims of this study were to (1) estimate the lifetime prevalence
of patients who receive positive screen results for bipolar disorder in an
urban general medicine clinic; (2) compare demographic, clinical, and treatment
characteristics of patients who screen positive for bipolar disorder with
those who do not; (3) report on health functioning and impairment of screen-positive
patients; and (4) determine whether the primary care physician was aware of
a history of manic symptoms at the time of screening.
The study was conducted at the faculty and resident group practice of
the Division of General Medicine, Columbia University Medical Center in New
York City between December 2001 and January 2003. The practice serves approximately
18 000 adult patients from the surrounding community each year.
The institutional review boards of the Columbia University Medical Center
and the New York State Psychiatric Institute approved the study protocol.
All participants provided written informed consent.
A systematic sample of consecutive adult patients seeking primary care
at the practice was invited to participate. Patients were systematically approached
to determine their eligibility on the basis of the position of the seat they
freely selected in the waiting room. Every consecutive patient from the chairs
in the back of the room to the front were screened for eligibility to obtain
our final goal of about 1000 patients. Eligible patients were between 18 and
70 years of age, had made at least 1 prior visit, could speak and understand
English or Spanish, and were waiting for a scheduled face-to-face contact
with their primary care physician. Patients were excluded if their current
general health status prohibited completion of the survey form.
Because one aim of the study was to examine primary care detection and
management of patients who screened positive for bipolar disorder, we limited
the sample to returning patients, as these patients are likely to be better
known to the primary care physicians than patients making their first clinic
visit. We also excluded a substantial number of waiting room patients who
were scheduled to see other health care professionals, were picking up medications
but not seeing a physician, or who were persons accompanying patients.
A total of 3807 patients were approached, of whom 169 (4.4%) refused
to participate. Of the 3638 who were prescreened, 2291 (63.0%) were ineligible
to participate. Common criteria for exclusion were: (1) not being scheduled
for face-to-face contact with a primary care physician (56.5%); (2) not being
between 18 and 70 years of age (33.5%); and (3) not having made a previous
visit to the practice (16.7%). Less commonly, patients were excluded because
of poor physical health (3.3%) or cognitive impairment (1.6%). Of the 1347
who met eligibility criteria, 1157 (85.9%) consented to participate. A total
of 1146 patients (99.0%) completed the screen.
All data forms were translated from English to Spanish and back-translated
by a bilingual team of mental health professionals. All participants completed
a sociodemographic history form to assess sex, age, annual household income,
race/ethnicity, marital status, educational achievement, and occupational
Participants also completed the Mood Disorder Questionnaire (MDQ), a
15-item self-report assessment of lifetime bipolar disorder based on Diagnostic and Statistical Manual of Mental Disorders, Fourth
Edition(DSM-IV) criteria.28 Standard
MDQ scoring for bipolar disorder requires endorsement of at least 7 lifetime
manic symptoms, several co-occurring symptoms, and moderate or serious associated
functional impairment. With these criteria, the MDQ has a sensitivity of 0.28
and a specificity of 0.97 in a community sample29;
a sensitivity of 0.73 and a specificity of 0.90 in an outpatient psychiatric
sample28 in relation to Structured Clinical
Interview for DSM-IV bipolar I and bipolar II disorders.
In the current analysis, participants who met or exceeded the standard MDQ
scoring algorithm were considered to have screened positive for a lifetime
history of bipolar disorder. Patients were also asked the age at which symptoms
began to be a problem and whether they had consulted a health professional
about these symptoms.
The survey forms included the DSM-IV Primary
Care Evaluation of Mental Disorders (PRIME-MD) Patient Health Questionnaire30 to assess current symptoms of major depression, panic
disorder, general anxiety disorder, and past-year probable alcohol abuse/dependence.
In addition, past year probable drug abuse/dependence was assessed with a
module patterned after the Patient Health Questionnaire alcohol use disorder
module. Suicidal ideation was assessed by asking whether patients experienced
thoughts of being better off dead or of hurting themselves in some way during
the last 2 weeks.
Physical and mental health functioning were based, respectively, on
the Physical and Mental Component Summary scores of the Medical Outcomes Study
12-Item Short Form Health Survey (SF-12).31 Impairment
was evaluated with the 10-point self-rated social life and family life/home
responsibilities subscales of the Sheehan Disability Scale (0, none; 1-3,
mild; 4-6, moderate; 7-9, marked; 10, extreme).32 Significant
impairment for each subscale was defined by a rating of 7 or higher. Because
only 229 (20.0%) of the patients were gainfully employed, the work subscale
of the Sheehan Disability Scale was not used in the following analyses. Independent
of the record review (described in the next section), self-report information
was collected from the patient on mental health diagnoses given by a health
professional and on mental health treatment history, including psychotropic
The institutional review board required that study participants provide
specific written authorization to notify their physician about screening results;
8.7% of patients gave permission. If patients reported current suicidal ideation
or appeared acutely distressed, they were referred to the clinic psychiatrist
A medical record review of primary care visits was conducted by 2 board-certified
psychiatrists (A.F. and M.O.). International Classification
of Diseases, Ninth Revision diagnoses claimed for the index visit and
the preceding 6 months were used to assess whether the primary care physicians
were currently aware of the patient’s history of bipolar disorder symptoms.
All available medical records were reviewed for evidence that a primary care
physician recognized bipolar disorder or symptoms, depressive disorders or
symptoms, or other mental disorders or psychiatric symptoms during the index
visit and a 6-month period preceding that visit. Due to the potential confound
with somatic illness, sleep and appetite disturbance were not considered depressive
Descriptive methods were used to examine background characteristics
of the entire sample. A positive screen for lifetime bipolar disorder was
determined to estimate the prevalence of bipolar disorder by scoring positive
responses to MDQ items, and the 95% confidence interval (CI) was determined.
Rates of major depression, panic disorder, general anxiety disorder, and alcohol
and drug use disorders were based on diagnostic algorithms for the PRIME-MD
Patient Health Questionnaire.30
The prevalence of a positive screen for bipolar disorder was stratified
by age (18-44, 45-54, 55-64, and 65-70 years), sex, and race/ethnicity. Race/ethnicity
was assessed in order to evaluate possible health disparities. Patients were
categorized as Hispanic if they identified their national origin/ancestry
as Latin American, were born in a Spanish-speaking country, or if they chose
to complete the study forms in Spanish. Non-Hispanic patients classified themselves
as black, white, or other.
Comparisons between patients who did and did not screen positive for
bipolar disorder on categorical variables (past mental health diagnoses, current
mental conditions, mental health care, social and family life impairment,
and work loss) were made with the χ2 test, except when any
cell had an expected count of less than 5, in which case the Fisher exact
test was used. Student’s t test was used for
comparisons involving continuous variables (SF-12 mental and physical component
summary scores). The percentage of missing data in each group was less than
5% except where noted.
Logistic regression was used to measure the relative risk (RR) of the
various diagnostic, health-related, and treatment characteristics as a function
of a positive screen for bipolar disorder. Because many of these characteristics
were relatively common (>10%), and odds ratios (ORs) are known to overestimate
RRs in such circumstances,33 ORs from the logistic
regression output were converted to RRs.34 Linear
regression was used to determine the expected change in SF-12 mental and physical
component summary scores due to a positive screen for a history of bipolar
disorder. For those who screened positive for bipolar disorder, we examined
past-month use of psychotropic prescriptions.
Of 1146 participants who completed the screen, 796 (69.5%) were female
and the mean (SD) age was 51.0 (12.2) years. Eight hundred fifty (75.2%) reported
that their annual family income was below $12 000. Nine hundred forty-one
(82.1%) were of Hispanic origin and of non-Hispanic participants, 149 (72.7%)
were black. Most participants (782; 68.4%) had never married or were currently
separated, divorced, or widowed. Four hundred nine (36.1%) had completed 8
or fewer years of education. Eleven hundred thirty-three (98.9%) had health
insurance. Participants reported whether they were paid workers (229; 20.0%),
unemployed (222; 19.4%), or described their occupational status as disabled
A total of 9.8% (95% CI, 8.1%-11.6%) of the participants had a positive
screen for a lifetime history of bipolar disorder. Among the 112 participants
who screened positive for bipolar disorder, the most commonly reported manic
symptoms were being very irritable (100; 89.3%) or “hyper” (99;
88.4%). Other frequently endorsed symptoms included being easily distractible
(98; 87.5%), having racing thoughts (97; 87.4%), and being more talkative
(96; 85.7%). Among those with a positive screen for bipolar disorder, the
least endorsed items were being more social or outgoing (59; 53.2%), being
more interested in sex (54; 48.2%), and having a decreased need for sleep
Participants with a positive screen for bipolar disorder first recognized
these symptoms as a problem at a mean (SD) age of 35.0 (13.1) years. Twelve
participants (11.3%) reported an onset at age 18 or younger, whereas 44 (41.5%)
had an onset at age 40 or older. Eighty-one participants (72.3%) who screened
positive for bipolar disorder had sought professional help specifically for
The prevalence of screening positive for bipolar disorder did not vary
significantly by sex, age, race/ethnicity, marital status, or level of educational
achievement (Table 1). Screening positive
for a history of bipolar disorder was significantly related to level of annual
household income, with the highest rates among the poorest respondents.
Nearly nine out of ten (88.4%) participants who screened positive for
bipolar disorder reported being previously diagnosed with a mental disorder
by a health professional, but only 9 (8.4%) reported having received a diagnosis
of bipolar disorder or manic depression (Table
2). Most commonly, these participants had been told that they had
depression (79.5%), or either anxiety/bad nerves or nervous breakdown (76.8%).
Twenty-two (19.6%) were diagnosed with an alcohol or drug use problem. The
patient’s report of any past mental diagnosis by a health professional
was strongly associated (ie, odds ratio >3.0) with having a positive screen
for bipolar disorder.
Nearly half (47.3%) of the participants who screened positive for bipolar
disorder met criteria for a current major depression episode (Table 2). Of all participants presenting with major depression,
nearly one quarter (23.5%) endorsed bipolar screen criteria. After adjustment
for demographic covariates, there was a strong association between screening
positive for bipolar disorder and current major depression.
Approximately half (48.1%) of the participants who screened positive
for bipolar disorder also met PRIME-MD DSM-IV criteria
for one or more anxiety or substance use disorders (Table 2). The most frequent comorbid disorder was generalized anxiety
disorder (25.2%), followed by alcohol use disorder (21.5%), panic disorder
(11.9%), and drug use disorder (9.1%). Among those who had current anxiety
or substance use disorder, the prevalence of screening positive for lifetime
bipolar disorder ranged from 23.3% (general anxiety disorder) to 32.3% (drug
use disorder). After adjustment for demographic covariates, screening positive
for a history of bipolar disorder was strongly associated with each of the
anxiety and substance use disorders.
Nearly one fifth (18.8%) of those who screened positive for bipolar
disorder as compared with 3.9% of those who screened negative reported suicidal
ideation at least some days during the previous 2 weeks. After controlling
for the presence of any current mental condition (major depression, panic
disorder, generalized anxiety disorder, alcohol use disorder, or drug use
disorder) and demographic covariates, screening positive for bipolar disorder
remained significantly associated with current suicidal ideation (RR = 4.80;
95% CI, 2.92-7.49).
Participants who screened positive for bipolar disorder were more likely
than those who screened negative to have had past mental health treatment
(Table 2). In logistic regression models
that adjusted for demographic covariates, screening positive for bipolar disorder
remained strongly associated with past use of prescribed psychotropic medications
and previous mental health hospitalization. Among patients who screened positive
for bipolar disorder, 43.9% (47/107) reported that they took a prescribed
psychotropic medication within the past month, but only 6.5% (7/107) reported
taking a mood-stabilizing medication such as lithium, valproate, or carbamazepine
in the past month.
Mental and physical health-related quality of life based on the SF-12
mental and physical component summaries was lower (worse) for those who screened
positive than negative for bipolar disorder (Table 3). These measures remained lower after controlling for relevant
covariates. Significant social and family life impairment were more common
among participants who screened positive than negative for bipolar disorder
(Table 3). Impairment in both areas
remained associated with screening positive for bipolar disorder in the multivariate
models. Work loss of 1 week or more in the past month was more commonly reported
by patients who screened positive than negative for bipolar disorder, but
was not significantly associated with a positive bipolar disorder screen after
controlling for the covariates.
Of 112 participants who screened positive for lifetime bipolar disorder,
100 had identifiable medical record numbers, and among these, 96 included
notes for 1 or more primary care visits during the 6-month review period.
A mental disorder or psychiatric symptom was recorded in the records
of 67.7% of the patients who screened positive for bipolar disorder. Depressive
disorders or symptoms were noted in 55 (49.0%) of these medical records, but
none included an indication of a bipolar disorder diagnosis. Mood stabilizers
were prescribed by the primary care physicians in 7.3% (7/96) of the cases.
No International Classification of Diseases, Ninth Revision codes for bipolar disorder were found for any patients, at any visits,
during the 6-month review period. During the period of this study, primary
care physicians at this practice did not have access to past psychiatric treatment
records, unless they specifically requested them from the facility providing
Nearly 10% of participants at an urban general medicine clinic screened
positive for lifetime bipolar disorder. This is one of the highest reported
lifetime estimates of the rates of bipolar disorders in primary care. Past
primary care studies have estimated the rate of bipolar disorders between
0.7% and 1.2%.6,11,14 These
studies used diagnostic instruments that may have poor sensitivity for detecting
hypomanic episodes and thus may underestimate bipolar II disorder.35 Bipolar II disorder, characterized by recurrent depressive
episodes and brief periods of hypomania, may be the most prevalent form of
Several national and community studies37-41 have
estimated the prevalence of a broader spectrum of bipolar disorders. These
studies have found a lifetime prevalence between 2.6% and 6.5%. A recent study
of a large nationally representative community sample reported a lifetime
prevalence of 3.7% with the MDQ.42
The high estimated prevalence in this clinical setting (9.8%) may be
related to the low socioeconomic status of the population. In a national study,42 lifetime prevalence of bipolar disorder was highest
(5.7%) among participants with the lowest annual household income (<$20 000/y).
In our clinical sample, nearly nine in ten participants reported a household
income below $18 000 per year, and the rate of screening positive for
lifetime bipolar disorder was inversely associated with household income.
These findings are consistent with community-based studies43,44 that
have shown that economically disadvantaged individuals have higher rates of
mental disorders than their more affluent counterparts. Previous research45,46 at our clinic site has also found
high rates of major depression, psychotic symptoms, and suicidal ideation.
The reported mean age of onset of bipolar disorder (35 years) is approximately
a decade older than reported in community samples. The later onset in this
sample may be due to the problem of recall in an older primary care sample
or may reflect a true later age of onset in this sample.
Remarkably few participants who screened positive for bipolar disorder
reported that they had been told by a health professional that they had a
bipolar disorder, though a majority had sought professional help for these
symptoms and had been previously diagnosed with a mental health condition.
Several factors may explain the apparent low rate of professional diagnosis
of bipolar disorders. First, participants may not recall receiving the diagnosis
or may not have understood the diagnosis. Second, assessment and treatment
by mental health specialists may not reduce misdiagnosis of bipolar disorder.47 Third, two fifths of participants who screened positive
for bipolar disorder reported a late onset (age 40 or older). Late-onset bipolar
disorder has been associated with a less severe symptom course than early
onset bipolar disorder48 and, as a result,
may be more likely to pass undiagnosed. Last, the racial and ethnic composition
of our sample may place them at a particularly high risk of being misdiagnosed.49
A majority of participants with a history of manic symptoms in our study
had current symptoms of major depression, anxiety, or substance use disorders.
Nearly half screened positive for major depression, a finding similar to those
of 2 studies that found that 31.9% of participants with bipolar I disorder50 and 50.3% of participants with bipolar II disorder51 experienced depressive symptoms during weekly mood
ratings. The high rates of anxiety and substance use disorders among participants
with lifetime manic symptoms are consistent with community and clinical studies.6,52
Screening low-income primary care patients with current mental conditions
can improve the detection of bipolar disorders. Poor people are less likely
to receive mental health treatment53 and, if
they do, are comparatively less likely to receive treatment from a mental
health specialist.54,55 Because
the poor tend to rely disproportionately on primary care physicians for mental
health treatment,56 these health professionals
can play an important role in the assessment and management of bipolar disorders.
Current guidelines for bipolar disorders27 caution
against monotherapy with antidepressants since these agents may induce a hypomanic,
manic, or mixed depressive/manic episode.25,26 In
our study, almost two thirds of participants with lifetime manic symptoms
who received medication in the past month reported taking antidepressant monotherapy.
In comparison, one third of participants with bipolar disorders who are seen
by outpatient psychiatrists have been found to take pharmacotherapy without
a mood stabilizer.57 Given the recent increase
in the prescribing of antidepressants,58 there
is a growing risk that misdiagnosed cases of bipolar disorder in primary care
may receive inappropriate treatment.
Bipolar disorders are associated with significant disability.59 Many patients with these disorders do not fully recover
ability to function in work and social activities60 and
they remain impaired despite maintenance treatment.61-65 Our
study demonstrates that primary care patients who screen positive for bipolar
disorder also experience significant disability in health, social, family,
and occupational functioning. Even after adjusting for the presence of any
current mental condition, impairment in health-related quality of life, social
activities, and in family life remained significantly associated with a positive
screen for bipolar disorder.
Participants with bipolar disorders are at high risk for suicide attempts
and completion.66-70 Our
study indicates that primary care participants who screen positive for bipolar
disorder are at increased risk of current suicidal ideation, even after controlling
for several other current mental conditions. Screening primary care patients
for bipolar disorder should also include ongoing assessment and management
of suicide risk. Given the high rates of morbidity and mortality associated
with bipolar disorders, detected primary care patients may be better referred
to mental health specialists who can provide regular visits and who are trained
to use complex pharmacologic regimens and manage acute crises.
The current study has several limitations. First, we used a self-report
instrument, the MDQ, to screen for bipolar disorder. The MDQ may be less accurate
than a structured diagnostic interview undertaken by a health professional
and may overestimate the lifetime prevalence of bipolar disorder. Second,
the specificity and sensitivity of the MDQ have not yet been established in
a primary care setting. The instrument’s specificity has been shown
to be high (0.97) in a large national sample,42 but
may be lower in our study setting because manic symptoms may be caused by
medical illnesses, such as hyperthyroidism, and medications, such as corticosteroids.
Its sensitivity may have been modest in our sample, given that only about
one out of five participants who were given a past diagnosis of bipolar disorder
or manic depression screened positive. Third, though we asked patients whether
they had ever been diagnosed with bipolar disorder or manic depression, we
did not ask specifically about other related conditions, such as manic episode,
hypomanic episode or mania that might have uncovered additional information.
Last, because the study was undertaken in an urban general medical practice
serving a largely low-income population, the findings may not generalize to
primary care settings with different populations.71,72
In an urban general medicine practice, screening positive for bipolar
disorder is relatively common but frequently underrecognized and is associated
with poor health-related quality of life, impairment in social activities
and family life, and current suicidal ideation. A significant proportion of
primary care participants who screened positive for bipolar disorder present
with major depression or an anxiety or substance use disorder. These participants
are at risk for adverse events if prescribed antidepressant monotherapy. However,
addition of a selective serotonin reuptake inhibitor does not appear to increase
the risk of switching from depression to mania in patients concurrently treated
with a mood stabilizer or antipsychotic medication.73 To
improve the recognition and reduce the morbidity of bipolar disorders in primary
care, further efforts are needed by primary care physicians to screen selectively
for past hypomania or mania among participants with known depression, anxiety,
or substance use conditions.
Corresponding Author: Myrna M. Weissman,
PhD, Department of Psychiatry, Columbia University, New York State Psychiatric
Institute, 1051 Riverside Dr, Unit 24, New York, NY 10032 (firstname.lastname@example.org).
Author Contributions: Dr Weissman had full
access to all of the data in the study and takes responsibility for the integrity
of the data and the accuracy of the data analysis.
Study concept and design: Das, Olfson, Pilowsky,
Feder, Gross, Shea, Weissman.
Acquisition of data: Das, Lantigua, Shea, Weissman.
Analysis and interpretation of data: Das, Olfson,
Gameroff, Blanco, Gross, Neria, Shea, Weissman.
Drafting of the manuscript: Das, Olfson, Gameroff,
Pilowsky, Gross, Weissman.
Critical revision of the manuscript for important
intellectual content: Das, Olfson, Pilowsky, Blanco, Feder, Neria,
Lantigua, Shea, Weissman.
Statistical analysis: Das, Gameroff, Shea.
Obtained funding: Olfson, Weissman.
Administrative, technical, or material support:
Das, Feder, Lantigua, Weissman.
Study supervision: Olfson, Pilowsky, Shea,
Financial Disclosure: None reported.
Funding/Support: This project was supported
by an investigator-initiated grant from Eli Lilly & Co (Dr Weissman),
and a National Research Service Award Institutional Research Training Grant
from the National Institute of Mental Health (Dr Das).
Role of the Sponsor: Neither agency had a role
in the study design or conduct; data collection, management, analysis, or
interpretation; or manuscript preparation, review, or approval.