Customize your JAMA Network experience by selecting one or more topics from the list below.
Olfson M, Marcus SC, Druss B, Elinson L, Tanielian T, Pincus HA. National Trends in the Outpatient Treatment of Depression. JAMA. 2002;287(2):203–209. doi:10.1001/jama.287.2.203
Context Recent advances in pharmacotherapy and changing health care environments
have focused increased attention on trends in outpatient treatment of depression.
Objective To compare trends in outpatient treatment of depressive disorders in
the United States in 1987 and 1997.
Design and Setting Analysis of service utilization data from 2 nationally representative
surveys of the US general population, the 1987 National Medical Expenditure
Survey (N = 34 459) and the 1997 Medical Expenditure Panel Survey (N
= 32 636).
Participants Respondents who reported making 1 or more outpatient visits for treatment
of depression during that calendar year.
Main Outcome Measures Rate of treatment, psychotropic medication use, psychotherapy, number
of outpatient treatment visits, type of health care professional, and source
Results The rate of outpatient treatment for depression increased from 0.73
per 100 persons in 1987 to 2.33 in 1997 (P<.001).
The proportion of treated individuals who used antidepressant medications
increased from 37.3% to 74.5% (P<.001), whereas
the proportion who received psychotherapy declined (71.1% vs 60.2%, P = .006). The mean number of depression treatment visits
per user declined from 12.6 to 8.7 per year (P =
.05). An increasingly large proportion of patients were treated by physicians
for their condition (68.9% vs 87.3%, P<.001),
and treatment costs were more often covered by third-party payers (39.3% to
Conclusions Between 1987 and 1997, there was a marked increase in the proportion
of the population who received outpatient treatment for depression. Treatment
became characterized by greater involvement of physicians, greater use of
psychotropic medications, and expanding availability of third-party payment,
but fewer outpatient visits and less use of psychotherapy. These changes coincided
with the advent of better-tolerated antidepressants, increased penetration
of managed care, and the development of rapid and efficient procedures for
diagnosing depression in clinical practice.
Depressive disorders are highly prevalent in the United States.1,2 Results from 2 large community-based
mental health surveys, the National Institute of Mental Health Epidemiologic
Catchment Area (ECA) survey (1980-1982) and the National Comorbidity Survey
(NCS) (1990-1992), suggest that the 1-year prevalence of major depression
in the adult population is between 5.0%1 and
10.3%.2 Cross-national epidemiologic research
further suggests that major depression is common in Europe, Canada, New Zealand,
and, to a lesser extent, Taiwan and Korea.3
Depressive disorders often impair social, occupational, and role functions.4-6 The detrimental effects
of depression on quality of life and daily function match those of heart disease
and exceed those of diabetes, arthritis, and peptic ulcer disease.7 According to the Global Burden of Disease Study,8 unipolar major depression is the fourth leading cause
of worldwide disability and is expected to become the second leading cause
Controlled clinical trials demonstrate that antidepressants and some
psychotherapies significantly reduce the symptoms of depression.9-11
Antidepressant medications are clinically effective across the full range
of severity of major depressive disorders.12,13
In addition, specific forms of time-limited psychotherapy are as effective
as antidepressants for mild to moderate depressions.13,14
Most individuals with depression receive no treatment for their symptoms.1,15,16 According to a recent
report of the surgeon general, promoting treatment for people with depression
is an even more significant problem than developing more effective treatments.17 In comparison to the extensive literature on the
efficacy of psychotherapy and pharmacologic treatments, remarkably little
is known about access to treatment for depression and the treatment experiences
of those who gain access. The late 1980s and first half of the 1990s was a
period of significant change in the delivery of mental health services, including
the growth of managed care and the development of selective serotonin reuptake
inhibitor (SSRI) medications. Given these changes, there is a dearth of information
regarding changes over time in characteristics of persons treated for depression.
In this article, we examine national trends in the care of outpatients
with depression using data from the 1987 National Medical Expenditure Survey
(NMES) and the 1997 Medical Expenditure Panel Survey (MEPS). These surveys
provide large, nationally representative samples and use methods that permit
comparisons to be made across the 2 points in time.
Data were drawn from the household component of the 1987 NMES18 and 1997 MEPS.19 Both
surveys were sponsored by the Agency for Healthcare Research and Quality to
provide national estimates of the use, expenditures, and financing of health
services. The NMES and MEPS were conducted as national probability samples
of the US civilian, noninstitutionalized population and were designed to provide
nationally representative estimates to be compared over time.
The 1987 NMES used a sampling design in which 15 590 households
were selected from within 165 geographic regions across the United States.
A sample of 34 459 individuals was included in the study, representing
a response rate of 80.1%. The 1997 MEPS household component was drawn from
a nationally representative subsample of the 1995 National Health Interview
Survey, which used a sampling design similar to that of the 1987 NMES. A sample
of 32 636 participants from 14 147 households was interviewed. This
represents a 74.1% response rate. For both surveys, a designated informant
was queried about all related persons who lived in the household.
The Agency for Healthcare Research and Quality devised weights to adjust
for the complex survey design and yield unbiased national estimates. The sampling
weights also adjust for nonresponse and poststratification to population totals
based on US census data. More complete discussions of the design, sampling,
and adjustment methods are presented elsewhere.19,20
Households selected for the 1987 NMES household survey were interviewed
4 times to obtain health care utilization information for the 1987 calendar
year.18 The 1997 MEPS included a series of
3 in-person interviews for 1997.19 In both
surveys, respondents were asked to record medical events as they occurred
in a calendar or diary, which was reviewed in-person during each interview.
Written permission was obtained from survey participants to contact medical
practitioners they or household members reported seeing during the survey
period to verify service use, medications, charges, and sources and amounts
of payment. Verification procedures were implemented for all pharmacy purchases,
health maintenance organization visits, and outpatient hospital visits and
for half of office-based visits.
Respondents were asked the reason for every outpatient visit during
the reference period. Conditions were recorded by interviewers as verbatim
text and then subsequently coded by professional coders according to the International Classification of Diseases, Ninth Revision
(ICD-9), as revised for the National Health Interview
Survey.21 Interviewers each underwent 80 hours
of training, and coders all had degrees in nursing or medical record administration.
A total of 5% of records were rechecked for errors; error rates in these rechecks
were less than 2.5%. A staff psychiatric nurse established mental disorder
diagnoses in cases of diagnostic ambiguity or uncertainty. Respondents who
made 1 or more outpatient visits coded for the purpose of major depressive
disorder, single episode (Diagnostic and Statistical Manual
of Mental Disorders, Fourth Edition [DSM-IV]/ICD-9 code 296.2); major depressive disorder, recurrent
(DSM-IV/ICD-9 code 296.3);
dysthymic disorder (DSM-IV/ICD-9 code 300.4); or depressive disorder, not otherwise specified (DSM-IV/ICD-9 code 311) were defined
as having received treatment for depression.
The MEPS and NMES booklets solicit information on the type of health
care professionals providing treatment at each visit. We classified health
care professionals into the following groups: physicians of all specialties
(a breakdown by physician specialty was not available for 1997), social workers,
psychologists, and a residual group of other providers that included nurses,
nurse practitioners, physician assistants, chiropractors, and other health
The NMES and MEPS asked respondents the type of care provided during
each outpatient visit using a flash card with various response categories.
Visits that included a specific indication that "psychotherapy/mental health
counseling" was provided are considered psychotherapy visits.
The NMES and MEPS asked for all prescribed medicines associated with
each health care visit. Respondents were asked to supply the names of any
prescribed medications purchased or otherwise obtained, the first and last
dates taken, the number of times obtained, and the conditions associated with
each medicine. We focus on prescribed psychotropic medications associated
with visits for the treatment of depression. Psychotropic medications were
classified as antidepressants, anxiolytics, antipsychotics, mood stabilizers,
or stimulants according to the 1997 Physicians' Desk Reference.22 A subcategory of antidepressants
was created for SSRIs that included fluoxetine hydrochloride, sertraline hydrochloride,
paroxetine, and fluvoxamine maleate. A subcategory of newer antipsychotics
included clozapine, olanzapine, and risperidone.
Rates of treatment for depression per 100 persons for each survey year
were computed stratified by sociodemographic characteristics. We then examined
sociodemographic characteristics of respondents who reported 1 or more health
care visits for depression in either survey. The χ2 test was
used to examine the strength of association between rates of treatment of
depression within sociodemographic categories and across survey years. Wald
F tests were used to identify differences in means of continuous variables
between the 2 survey years. An examination was also made of treatment characteristics
of individuals ("patients") who reported receiving treatment for depression
in each survey year.
We used a logistic regression model to evaluate the association between
survey year and psychotropic medication prescription. To adjust for changes
in patient characteristics between the survey years, we controlled for respondent
age, sex, race, marital status, education, employment status, and insurance
status. Logistic regression models were also used to estimate the effect of
survey year on rate of psychotherapy, antidepressant medication, and the combination
of psychotherapy and medication treatment. A multiple linear regression model
was used to evaluate the association between survey year and number of psychotherapy
visits, controlling for the various sociodemographic covariates. All statistical
analyses were performed using the SUDAAN software package23
to accommodate the complex sample design and the weighting of observations.
The α value was set at .05 and all tests were 2-sided.
Between 1987 and 1997, there was a significant increase in the overall
rate of outpatient treatment of depression. The rate of treatment increased
from 0.73 per 100 persons in 1987 to 2.33 in 1997 (Table 1). In contrast, there was little change during the study
period in the rate within the general population of persons who received any
outpatient general medical treatment (74.1% in 1987 and 72.4% in 1997).
A significant increase in the rate of outpatient treatment of depression
occurred across all sociodemographic groups examined. As a proportion of the
baseline rate of treatment, Hispanic and black persons experienced slightly
larger increases than white persons (Table
1). However, the rate of outpatient treatment for Hispanics and
blacks remained well below the rate of whites. The highest rates of treatment
were for divorced, separated, or widowed individuals, those with at least
a high school education, and unemployed persons (Table 1).
In both survey years, most patients who received outpatient treatment
for depression were between 18 and 64 years of age, white, female, employed,
and privately insured. Slightly less than half were married or had more than
a high school education (Table 2).
There was a significant increase in the rate of outpatient treatment
for depression regardless of insurance status. However, the rate of treatment
among individuals without insurance remained below that of individuals with
either private or public insurance (Table
The proportion of individuals treated for depression who received a
prescribed psychotropic medication increased from 44.6% in 1987 to 79.4% in
1997 (Table 3). After controlling
for the possible confounding effects of sociodemographic characteristics,
individuals treated for depression were 4.5 times more likely to be treated
with a psychotropic medication in 1997 than in 1987 (Table 4). During this period, the proportion of pharmacy costs paid
by third-party payers increased from 39.3% to 55.2% (Table 3).
Antidepressants were the most commonly prescribed medications for the
treatment of depression. After adjusting for confounding sociodemographic
factors, patients treated for depression were 4.8 times more likely to receive
an antidepressant in 1997 than in 1987 (Table 4). The increase in antidepressant use was primarily attributable
to SSRIs, a class of antidepressant medication that was unavailable in 1987.
Selective serotonin reuptake inhibitors were prescribed to more than half
(58.3%) of individuals who received outpatient treatment for depression in
Anxiolytics were the second most commonly prescribed class of psychotropic
medication in both survey years, but were prescribed to fewer than 1 in 7
patients treated for depression. Anxiolytics were followed by mood stabilizers
and antipsychotics. Stimulants were rarely prescribed (Table 3).
With the increase in rate of outpatient treatment for depression, there
was a corresponding increase in the rate of psychotherapy for depression.
However, among persons treated for depression, the percentage who received
psychotherapy declined from 71.1% (1987) to 60.2% (1997) (Table 3). Among those who received psychotherapy, the mean annual
number of psychotherapy visits declined from 12.6 visits in 1987 to 8.7 visits
in 1997. This decline remained statistically significant after controlling
for the effects of patient sociodemographic characteristics (Table 4). During this period, there was an increase in the proportion
of psychotherapy costs borne by third-party payers (Table 3).
The proportion of patients treated for depression who received at least
1 psychotherapy visit along with a prescription for a psychotropic medication
increased from 28.8% in 1987 to 48.1% in 1997. The comparable proportions
who received at least 1 psychotherapy visit and a prescription for an antidepressant
were 23.2% in 1987 and 45.2% in 1997. Treated patients in 1997 were almost
twice as likely to receive psychotherapy and a psychotropic medication than
they were in 1987 after controlling for confounding sociodemographic factors
(odds ratio, 2.0) (Table 4). They
were also 2.4 times more likely to receive psychotherapy and an antidepressant
During the study period, there was a significant increase in the proportion
of patients whose treatment of depression involved visits to a physician (Table 3). By 1997, more than 8 (87.3%)
of 10 patients who received outpatient treatment of depression were treated
by a physician compared with 68.9% in 1987. Conversely, the percentage who
received treatment from psychologists declined (29.8% vs 19.1%). Treatment
of depression by social workers remained little changed and relatively uncommon.
Significant growth occurred in the number of Americans who received
treatment for depression during the past decade, and at the same time the
treatments they received underwent a profound transformation. Antidepressant
medications became established as a mainstay, psychotherapy sessions became
less common and fewer among those receiving treatment, and physicians assumed
a more prominent role. These changes suggest that access to mental health
services has increased and that there has been an increased emphasis on pharmacologic
Several factors may have contributed to these trends. Beginning with
the introduction of fluoxetine in late 1987 and followed by several other
SSRIs and antidepressants with atypical mechanisms of action, there has been
a steady broadening of the pharmacologic options available to treat depression.
The new medications tend to have fewer adverse effects,24,25
require less complicated dosing regimens, and pose less danger when taken
in overdose than the older tricyclic antidepressants. The comparative safety
and ease of prescribing SSRIs and the other newer antidepressants may have
led physicians to lower the symptom severity threshold at which they decide
to prescribe an antidepressant.26 If the availability
of the newer medications tipped the balance in favor of diagnosing and treating
depression, this would help explain both the increase in the overall rate
of treatment and the increase in the proportion of treated cases who filled
prescriptions for antidepressant medications.
The pharmaceutical industry also engaged in a concerted effort to promote
the increased sale of these new antidepressant medications through vigorous
advertising campaigns directed at physicians, other health care professionals,
and more recently the general public.27,28
In addition, medications to treat depression have been a featured topic of
lead articles in national news magazines, best-selling books, and widely watched
television talk shows. A new generation of screening and diagnostic instruments,
developed through partnerships between industry and academia, has also become
available to facilitate the rapid and efficient detection of depression in
Beginning in late 1987, the federal government embarked on a public
health campaign to educate the public and the medical community about the
recognition and treatment of depression.32
In 1991, the National Depression Screening Day program was inaugurated to
increase awareness and treatment of depression. By 1997, there were more than
2800 screening sites in the 50 states and Canada.33
These campaigns have underscored the importance of pharmacologic treatments.
In addition, there have been institutional efforts to improve the diagnosis
of depression and influence physician prescribing practices through the publication
of treatment guidelines.12,13
As a result of these developments, the public may have become more accepting
of pharmacologic treatment of depression. According to a 1986 Roper poll,34 only 12% of respondents indicated that they would
be willing to take medication for depression, whereas 78% stated they would
live with the depression until it passed. An ABC News poll35
conducted in April 2000 found that 28% of adults would be willing to take
antidepressants for depression for an extended period even though they were
informed that safety studies had not been conducted on long-term use of these
medications. This suggests that the pharmacologic treatment of depression
is becoming less stigmatized.
The growth in managed mental health care and the concepts of disease
management36 and medical necessity37 may have further spurred the pharmacologic treatment
of depression. In many plans, comprehensive pharmacy benefits encourage medication
management visits over psychotherapy visits, which are not reimbursed as generously.
Managed care generally seeks to shift patient care from specialty to primary
care physicians who are able and may be more likely to use pharmacologic treatments
rather than psychotherapy to manage depression. In addition, mental disorders
that require ongoing treatment are increasingly managed in behavioral health
care "carve outs" that seek to reduce costs by lowering the number of visits
per depressed individual.38,39
The comparatively low rate of treatment among black and Hispanic individuals,
those with less education, and those without health insurance suggests that
an unmet need for treatment may be especially great within these groups. Epidemiologic
data indicate that the rate of major depression is inversely related to income
and educational achievement and that depression is more common among Hispanics
than whites or blacks.1,40 These
findings suggest, but cannot confirm, that individuals within these minority
groups are vulnerable to undertreatment.
According to the NCS (1990-1992), 3.1% of adults 15 to 54 years of age
received outpatient health care treatment in 1 year for a mood disorder.15 The earlier ECA (1980-1982) survey reported that
3.6% of adults have an affective disorder and receive mental health treatment
in the health system during 1 year.1 These
NCS and ECA findings exceed the corresponding findings from the NMES (1987)
(0.7%). This disparity may be related to important methodologic differences
between the studies. For example, although the ECA figures include treatment
for any mental health or addictive symptoms by adults who meet criteria for
an affective disorder, the NMES figures include only outpatient treatment
reported for depression. Methodologic differences between the studies stem
from underlying differences in their primary aims: to quantify psychopathology
in the community (ECA and NCS) and to measure service use over time (NMES
One consequence of increased pharmacologic treatment of depression is
that larger numbers of depressed individuals are being treated with both pharmacotherapy
and psychotherapy. Recent research suggests that the combination of an antidepressant
and cognitive behavioral psychotherapy is more efficacious than either treatment
alone for chronic forms of major depression.41
In milder depressions, psychotherapy alone may be nearly as effective as the
combination of antidepressants and psychotherapy.42
The extent to which combined treatments confer meaningful advantages over
single-modality treatments in clinical practice awaits detailed longitudinal
practice-based outcomes research.
The current study is constrained by several limitations. Both the NMES
and MEPS collect data from household informants who may not be fully aware
of all of the services used by household members. Stigma, recall problems,
and problems distinguishing the different provider groups pose threats to
the reporting and classification of the survey data. Some respondents, especially
those with less education, may not be able to identify "psychotherapy/mental
health counseling" when they receive it. Without an independent measure of
symptoms, it is not possible to determine whether patients who received treatment
actually met diagnostic criteria for the selected conditions. Finally, the
1997 survey did not break out providers by physician specialty, thereby limiting
our ability to determine whether the observed changes in treatment patterns
occurred primarily in the general medical or specialty mental health sector.
In recent years, a growing number of Americans have received treatment
for depression. During this period, antidepressant medications have gained
popularity and physicians have extended their involvement in care. For the
promise of increased access to treatment to be fully realized, available treatments
must be provided in a safe, timely, and effective manner. An important challenge
ahead is to characterize the community treatment of depression with greater
specificity and relate variations in these treatments to critical patient