P&T DIGEST

A PEER-REVIEWED COMPENDIUM OF FORMULARY CONSIDERATIONS

DEPRESSION
• Prevalence and
economic implications

• Guidelines for
treatment

• Drug-therapy review

• Pharmacoeconomics
of treatment

• Adherence to therapy

• NCQA/HEDIS
measures

• Social anxiety disorder

Continuing education credit for physicians and pharmacists

sponsored by The Chatham Institute

Supported by an educational grant from

 A B O U T T H I S P U B L I C AT I O N

P&T DIGEST
A PEER-REVIEWED COMPENDIUM OF FORMULARY CONSIDERATIONS

Chief Medical Editor

David V. Sheehan, MD, MBA
Faculty
Christy L. Beaudin, PhD, LCSW, CPHQ
Bernard S. Bloom, PhD
Robert Burchuk, MD
Laura E. Frankum, PharmD
A Tool for Formulary Decision Makers
Benjamin G. Druss, MD, MPH

D
epression is one of the most common illnesses in the
Vincent J. Giannetti, PhD United States and a major driver of health care utilization
Ashok Raj, MD
costs. The literature is replete with studies that show that
Matthew W. Sarnes, PharmD
Jeffrey B.Weilburg, MD depression is managed poorly and that failure to diagnose and
treat depression can lead to poor yet avoidable medical and fi-
Editor
nancial outcomes. The primary goals and challenges of treating
Michael D. Dalzell
depression involve accurate diagnosis of this highly comorbid
Consulting Editor condition and improving patient compliance with therapy.
John A. Marcille
This peer-reviewed publication is a digest of up-to-date guide-
Senior Science Editor lines for treatment, therapeutic approaches to treatment of de-
Paula R. Sirois pression, pharmacoeconomic considerations in treatment, and
Contributing Editors a discussion of comorbid conditions. In consolidating this in-
Tony Berberabe formation, it serves as a valuable tool for formulary committees
Frank Diamond and is an important contribution to the medical literature.
William W. Edelman
Linda Hull Felcone
Susan Keller
Sylvie Kestler Editorial Advisory Board
Maxine Losseff
Jack McCain David V. Sheehan, MD, MBA
Design Director Professor of Psychiatry
Philip Denlinger Director of Psychiatric Research
University of South Florida College of Medicine
Group Publisher
Timothy P. Search, RPh Herbert C. Schulberg, PhD
Director, New Product Development Professor of Psychology in Psychiatry
Timothy J. Stezzi Weill Medical College of Cornell University
Eastern Sales Manager Attending Psychologist in Psychiatry
Scott MacDonald New York Presbyterian Hospital, Cornell Campus
Senior Account Manager
Blake Rebisz Sabah N. Chammas, MD, MAS
Midwest Sales Manager President and CEO
Terry Hicks Psychiatric Centers at San Diego
Director, Production Services Medical Director
Waneta Peart Tri-City Medical Center Behavioral Health Unit
Circulation Manager
Jacquelyn Ott This publication is made possible by an educational
MANAGED CARE (ISSN 1062-3388) is published monthly by MediMedia
grant from
USA Inc. at 780 Township Line Road,Yardley, PA 19067.This is Volume 13,
Issue 6. Periodical postage paid at Morrisville, Pa., and at additional mail-
ing offices. POSTMASTER: Send address changes to MANAGED CARE,
780 Township Line Road,Yardley, PA 19067. Prices: $10 per copy, $93 per
year in the USA; $120 per year elsewhere. E-mail: editors_mail@
managedcaremag.com. Phone: (267) 685-2788; fax (267) 685-2966; cir-
culation inquiries (267) 685-2782. Copyright ©2004 MediMedia USA.
P&T DIGEST
A PEER-REVIEWED COMPENDIUM OF FORMULARY CONSIDERATIONS

DEPRESSION
David V. Sheehan, MD, MBA
Chief Medical Editor
P&T DIGEST
A PEER-REVIEWED COMPENDIUM OF FORMULARY CONSIDERATIONS

DEPRESSION
Continuing Education Objectives/Accreditation Statements....4
INTRODUCTION
Depression: Underdiagnosed, Undertreated,
Underappreciated ........................................................................6
Major depressive disorder is significantly underdiagnosed and undertreated, particularly
in primary care. Although more patients are seeking help for depression and utilization
of antidepressants is on the rise, treatment is inadequate. Rectifying this will involve pa-
tients, providers, payers, employers, accrediting agencies, and governmental entities.
DAVID V. SHEEHAN, MD, MBA

EPIDEMIOLOGY AND ECONOMICS

Prevalence and Economic Effects of Depression .......................9
The lifetime risk of major depression among Americans is 17 percent, with as many as 1
in 10 people suffering from depression in any 1-year period. The author reviews depres-
sion’s epidemiology, the costs of treatment and nontreatment, and its economic impact.
This review also examines ways to improve value for money spent relative to this disease.
BERNARD S. BLOOM, PHD

DIAGNOSIS AND TREATMENT

Clinical Practice Guidelines
for Treating Depression in Primary Care.................................17
Clinical practice guidelines for depression offer the best available treatment options, as
evidenced by randomized controlled clinical studies. Although the use of evidence-based
guidelines has been shown to improve treatment success, there are few data to suggest
the extent to which they are being used or their effect on treatment outcomes.
CHRISTY L. BEAUDIN, PHD, LCSW, CPHQ AND ROBERT BURCHUK, MD

PHARMACOTHERAPY
An Overview of SSRI and SNRI Therapies for Depression.....25
Selective serotonin reuptake inhibitors (SSRIs) and serotonin norepinephrine reuptake
inhibitors (SNRIs) are used widely to treat mood and anxiety disorders. Indications,
pharmacologic characteristics, dosing and administration, and clinical use are outlined.
JEFFREY B. WEILBURG, MD
COST-EFFECTIVENESS
Pharmacoeconomic Evaluation of Antidepressant Therapies....34
With today’s emphasis in the managed care environment on evidentiary literature, new
antidepressants must provide economic evidence of their value compared with previously
available agents. Various pharmacoeconomic models have shown newer antidepressant
agents to be cost-effective relative to older agents. The authors review five methodologies
of pharmacoeconomic analysis and published results about antidepressants.
MATTHEW W. SARNES, PHARMD, AND LAURA E. FRANKUM, PHARMD

COMPLIANCE
Adherence With Antidepressant Medication...........................42 COVER ART:
Nonadherence with antidepressant medications is a well-documented factor in treatment Simultaneous Windows
failure. This article uses the literature to build a collaborative counseling model to in- (2nd Motif, 1st Part),
Robert Delaunay, 1912.
crease antidepressant adherence. Studies may be warranted to determine whether specific Oil on canvas. Solomon
antidepressants, as part of a collaborative strategy, can improve long-term adherence. R. Guggenheim Museum.
VINCENT J. GIANNETTI, PHD © L&M Services B.V.,
Amsterdam.
QUALITY MEASUREMENT
A Review of HEDIS Measures and Performance
for Mental Disorders .................................................................48
Performance on mental health HEDIS measures has been modest, with only minimal im-
provements in recent years, and is consistently worse than HEDIS performance for other
conditions. A step toward improving performance is to understand where care is provided
and then identify clinicians responsible for delivering appropriate care.
BENJAMIN G. DRUSS, MD, MPH

COMORBIDITY
Overview and Treatment of Social Anxiety Disorder..............52
Social anxiety disorder is a highly prevalent disorder that is frequently comorbid with
major depressive disorder. Pharmacotherapies, including SSRIs and benzodiazepines, as
well as cognitive-behavioral interventions, are effective for many patients, helping them
attain a higher quality of life.
ASHOK RAJ, MD

Continuing Education Post-test .................................................58

CME/CPE Answer Sheet/Evaluation Form................................60
 S E L F - S T U DY CO N T I N U I N G E D U C AT I O N AC T I V I T Y
P&T DIGEST
Depression
Continuing education is offered to physicians and
pharmacists who read pages 6 through 57 of this publi-
cation, and complete the post-test and fill out the evalu-
ation form on page 60.
EDUCATIONAL NEEDS ASSESSMENT
Depression is a serious public health threat and, by
extension, a threat to the health of populations for
which managed care organizations have responsibility.
Medical and pharmacy directors and other decision
makers in managed care organizations seek state-of-
the-art information about evidence-based guidelines,
emerging therapies, and pharmacoeconomics.This pub-
lication serves as a digest of this information for man-
aged care decision makers, who can use it to develop
treatment protocols and formulary recommendations
and disseminate it to providers to improve the collective
health of their populations.
The information herein also is valuable to health care
providers, who must stay abreast of current treatment
approaches.The success of treatment depends on the
clinician’s ability to accurately diagnose patients and en-
courage adherence to therapy.This publication synthe-
sizes best-practice information for clinicians.
The subject matter in this publication was selected on
the basis of literature searches and faculty perception of
significant issues.
EDUCATIONAL OBJECTIVES
After reading this publication, the participant should
be able to:
• Illustrate the prevalence of depression, as well as
the clinical and economic implications thereof.
• Describe major published treatment guidelines for
depression, as well as their shortcomings as ad-
vanced therapies have reached the market.
• Identify various SSRI treatments for depression and
explain their mechanisms of action.
• Understand the dynamics of pharmacoeconomic
evaluation of SSRI therapy.
• Elucidate strategies for therapeutic compliance and
the consequences of noncompliance.
• Discuss NCQA guidelines for depression medica-
tion management and follow-up after hospitaliza-
tion for depression.
4 P&T DIGEST
TARGET AUDIENCES
Managed health care professionals, including physi-
cians, pharmacists, medical directors, chief medical offi-
cers, pharmacy directors, and other senior managers in
managed care organizations.
CONTINUING EDUCATION
MEDICAL ACCREDITATION
The Chatham Institute is accredited by the Accredita-
tion Council for Continuing Medical Education (ACCME)
to provide continuing medical education for physicians.
The Chatham Institute designates this educational ac-
tivity for a maximum of 3.0 category 1 credits toward
the AMA Physician’s Recognition Award. Each physician
should claim only those credits that he/she actually
spent in the activity.
This activity has been planned and implemented in
accordance with the ACCME Essential Areas, Elements,
and Policies.
PHARMACY ACCREDITATION
The Chatham Institute is approved by the American
Council on Pharmaceutical Education (ACPE) as a
provider of continuing pharmaceutical education.
This activity provides 3.0 contact hours (0.3
CEU) of continuing education for pharmacists.
Credit will be awarded upon successful com-
pletion of the post-test and the activity evaluation.
ACPE Universal Program Number (UPN):
812-999-04-009-H01.
Release date: June 15, 2004
Expiration date: June 15, 2005
Medium: Journal supplement
PLANNING COMMITTEE MEMBERS
Timothy P. Search, RPh, group publisher, MANAGED CARE,
a division of MediMedia USA; Michael D. Dalzell, editor,
custom publications, MediMedia USA Managed Markets
Publishing; Cyndi Grimes, managing director,The
Chatham Institute; David V. Sheehan, MD, MBA, professor
of psychiatry and director of psychiatric research, Uni-
versity of South Florida College of Medicine.
 S E L F - S T U DY CO N T I N U I N G E D U C AT I O N AC T I V I T Y
CONFLICT-OF-INTEREST POLICY AND DISCLO-
SURES OF SIGNIFICANT RELATIONSHIPS
As an accredited provider,The Chatham Institute re-
quires that its faculty comply with ACCME Standards
for Commercial Support of Continuing Medical Educa-
tion and disclose the existence of any significant finan-
cial interest or any other relationship a faculty member
may have with the manufacturer(s) of any commercial
product(s) or device(s). It also requires the faculty to
disclose discussion of off-label uses in their presenta-
tions.
FACULTY DISCLOSURES
Bernard S. Bloom, PhD, acknowledges a consulting
relationship with GlaxoSmithKline.
Sabah Chammas, MD, acknowledges consulting rela-
tionships with Janssen, GlaxoSmithKline, and Wyeth. He
has received honoraria from Pfizer,Wyeth, Janssen,
GlaxoSmithKline, Eli Lilly & Co., and AstraZeneca.
Laura E. Frankum, PharmD, acknowledges a consult-
ing relationship with GlaxoSmithKline.
Matthew W. Sarnes, PharmD, acknowledges a con-
sulting relationship with GlaxoSmithKline.
David V. Sheehan, MD, MBA, acknowledges grant and
research support from GlaxoSmithKline, as well as an
advisory and speakers bureau relationship with Glaxo-
SmithKline.
Jeffrey B.Weilburg, MD, acknowledges grant and re-
search support from Wyeth-Ayerst, Pfizer, and Eli Lilly &
Co. He also acknowledges a consulting relationship
with GlaxoSmithKline.
Christy L. Beaudin, PhD, LCSW, CPHQ; Robert Burchuk,
MD; Benjamin G. Druss, MD, MPH; Vincent J. Giannetti,
PhD; Ashok Raj, MD; and Herbert C. Schulberg, PhD, have
no financial interest, arrangement, or affiliation that
would constitute a conflict of interest concerning this
continuing education activity.
PUBLISHER’S STATEMENT
This publication is supported by an educational
grant from GlaxoSmithKline.The material in this publi-
cation has been independently peer reviewed.The
sponsor played no role in reviewer selection.
Opinions are those of the authors and do not neces-
sarily reflect those of the institutions that employ the
authors, GlaxoSmithKline, MediMedia USA, or the pub-
lisher, editor, or editorial board.
Clinical judgment must guide each clinician in
weighing the benefits of treatment against the risk of
toxicity. Dosages, indications, and methods of use for
products referred to in this publication may reflect the
clinical experience of the authors or may reflect the
professional literature or other clinical sources, and may
not be the same as indicated on the approved package
insert. Please consult the complete prescribing informa-
tion on any products mentioned in this publication.
MediMedia USA Inc. assumes no liability for any mater-
ial published herein.
P&T DIGEST 5
 INTRODUCTION INTRODUCTION
Depression: Underdiagnosed,
Undertreated, Underappreciated
DAVID V. SHEEHAN, MD, MBA1
University of South Florida
SUMMARY
Major depressive disorder is significantly un-
derdiagnosed and undertreated, particularly
in the primary care environment. Although
more patients are seeking help for depres-
sion and the utilization of antidepressants is
on the rise, the level of treatment is inade-
quate. Rectifying this will involve patients,
providers, payers, employers, accrediting
agencies, and even governmental entities.
M
ajor depressive disorder in the United States
is a serious, recurring, and debilitating ill-
ness. More than 16 percent of the popula-
tion, or approximately 35 million American
adults, will suffer from it at some point during their
lives. According to the National Comorbidity Survey
Replication, published last year in the Journal of the
American Medical Association, 12-month prevalence rates
for major depressive disorder (MDD) translate to 6.6 per-
cent of the population, or about 14 million U.S. adults
(Kessler 2003).
To illustrate the magnitude of depression, consider
this: In the United States, there were 30,622 suicides in
2001; most were by people with depression. By contrast,
there were 20,308 homicides the same year (CDC 2001).
Such widespread incidence and prevalence translates
to a significant deterioration of quality of life, medical well
being, and personal productivity for millions of Ameri-
1David V. Sheehan, MD, MBA, is professor of psychiatry at
the University of South Florida College of Medicine, pro-
fessor of psychology at the USF College of Arts and Sciences,
and director of the Psychiatric Research at USF College of
Medicine. He completed his residency in psychiatry at
Massachusetts General Hospital and Harvard Medical
School. At Harvard Medical School, where he was assistant
professor of psychiatry, he was on the full-time faculty for
13 years. He has written over 450 abstracts and over 200
publications, including a bestseller, The Anxiety Disease.
6 P&T DIGEST
cans. From a societal perspective, depression contributes
to substantial worker ab-
senteeism and disability and
erodes billions from the U.S.
economy. From a health
plan or employer perspec-
tive, MDD is one of the
most expensive disorders
that payers face.
MDD is not simply the ex-
perience of a few “blue days.”
It is characterized by a severe
persistent depressed mood
and loss of interest or plea-
sure in normal activities.
MDD commonly includes David V. Sheehan, MD, MBA
decreased energy, changes in
sleep patterns and appetite,
and feelings of guilt or hopelessness. To be classified as
MDD, these symptoms must be present for at least 2 weeks,
cause significant distress, and interfere with activities of
daily living. If the depression is extremely severe, it may be
accompanied by psychotic symptoms or by suicidal
thoughts or behaviors.
From all principal perspectives — the patient, the pro-
vider, and the payer — the problem is large and growing
larger. This is due, to a great extent, to the fact that MDD
is significantly underdiagnosed and undertreated — par-
ticularly in the primary care environment, where the pre-
ponderance of treatment for it takes place. Readers who
are responsible for reining in ever-increasing pharmacy
budgets may find this statement difficult to accept. With
antidepressant drug use becoming more common among
insured populations, their perspective has a certain va-
lidity. For example, the medical literature documents a
significant rise in the number and rate of outpatients
being treated for depression over the past 25 years. In
1987, 0.73 per 100 persons were treated on an outpatient
basis for depression; by 1997, that rate had increased to
2.33 per 100 persons — an increase of more than 300 per-
cent (Olfson 2002). The same trend is documented with
regard to antidepressant prescribing during office visits,

with a near doubling of such visits over approximately
the same time period (Olfson 1998).
So, while it is true that more patients are seeking help
for depression and that utilization of antidepressants is
on the rise, the question becomes: Is it adequate? The
simple answer is no. Referring to the JAMA article cited
earlier, it is clear that even with increased numbers of
people now receiving pharmaceutical treatment for de-
pression, this still only represents just over half of all pa-
tients diagnosed with the disease. Beyond that, of those
who are receiving treatment, less than 42 percent receive
adequate treatment. In short, less than 22 percent of all
persons diagnosed with depression receive adequate
treatment for their disease (Kessler 2003).
COLLECTIVE RESPONSIBILITY
There are numerous and complex reasons for this un-
fortunate situation. Rectifying it will involve patients,
providers, payers, employers, accrediting agencies, and
even governmental entities. We will touch on several of
these reasons in turn.
Anyone involved with any aspect of the care of de-
pressed patients must understand the recurrent and co-
morbid nature of the disease. The former serves to make
an accurate diagnosis of depression more likely; the lat-
ter frequently contributes to missed diagnosis or mis-
diagnosis. Depression often is accompanied by other
mental and physical comorbidities. These commonly in-
clude anxiety, substance abuse, heart disease, hyperten-
sion, diabetes, and chronic pain resulting from other
conditions. Designing an appropriate treatment plan in
the presence of comorbid conditions is understandably
a challenge for providers, particularly when the depres-
sion is the underlying disorder.
Stigma is still stubbornly associated with depression
and other mental illnesses. This must be addressed head-
on with educational campaigns aimed at providers of
health care, the public, and employers. The stigma of
mental illness is a barrier to good care: When patients are
unwilling or unable to share their symptoms with their
health care providers, an accurate diagnosis and a tar-
geted treatment plan will remain elusive.
Another obstacle to optimal care is lack of follow-up.
This can result from physician time constraints, a weak
link in the system that does not automatically schedule pa-
tients for an appropriate number of follow-up visits, or
ignorance on the part of providers that in-office visits or
even telephone consults can increase patient compliance
and improve clinical outcomes. According to numerous
studies, as well as anecdotal evidence, the well-informed
patient will be a more compliant partner.
This lack of adequate follow-up is seen in the primary
care setting as well as the specialist’s office. It also has been
INTRODUCTION
the National Committee for Quality Assurance’s point of
entry. NCQA is an accrediting body with a mission to
measure and improve the quality of health care in Amer-
ica. Using process measures as a proxy for quality, NCQA
has developed numerous indices to evaluate quality of
care. Several of these involve depression. Two of the mea-
sures concern adherence to medication regimens during
the acute (12-week) and continuation (6-month) stages
of the disease, and one measures the number of follow-
up visits with a primary care practitioner or mental
health care provider. NCQA has established that 3 visits
during the 12-week acute stage is the benchmark (1 visit
must be with a prescribing practitioner). Yet, few health
plans in America meet the benchmark; in 2002, across all
plans reporting, only 19.2 percent of patients with a new
episode of depression had at least 3 follow-up visits. Ob-
viously, systems are not in place to guarantee that patients
who are newly diagnosed with depression and prescribed
medication are seen for what some would argue is a
minimum necessary number of follow-up visits.
Besides lack of follow-up, dropout or noncompliance
with therapy is a major documented problem. In a well-
referenced study by Lin (1995), researchers found a 52
percent dropout rate by week 12 among a group of
Medicaid patients who were prescribed antidepressants;
only 27 percent remained on an adequate, 6-month
course of therapy. Again, several causative factors can be
identified. One is side effects; treatment-emergent nau-
sea is the most common. Yet, if patients are made aware
that this potential side effect is essentially gone within the
first 12 weeks of therapy, compliance rates may improve.
Another causative factor in dropout can be a per-
ceived lack of efficacy or unreasonable expectations
regarding the length of time needed for the antidepres-
sant to do its work. Again, good patient-provider com-
munication is critical if patients are to have a realistic pic-
ture of what to expect from therapy. Unfortunately, a
disconnect is evident in the respective perceptions of
providers and patients about the quality of their com-
munications. Providers assume that they have imparted
more and better information about the therapeutic reg-
imen than patients claim to have heard.
Yet another factor that may limit a patient’s compli-
ance is the cost of antidepressant medications. Employ-
ers and other payers are wise to give serious considera-
tion to health plans and insurance policies that provide
their employees, retirees, and dependents with adequate
medication coverage that makes treatment affordable.
Poor clinical outcomes resulting from lack of adherence
to a therapeutic regimen are avoidable, and the down-
stream costs of follow-up care are potentially much
higher than the cost of the medication itself. The most
scientifically robust, evidence-based treatment algorithm
DEPRESSION 7
 INTRODUCTION
will have little or no beneficial effect if patients cannot
or do not follow their medication protocols.
So, the challenges are many and multifaceted. They in-
volve people, organizations, policy, and process. Chang-
ing our nation’s poor performance in caring for persons
with major depression begins with awareness of these
problems. Awareness leads the problem-solver to seek in-
formation and answers. We hope that in the following
pages, we provide a foundation for taking steps to im-
prove both the process and the outcomes of care for the
millions of Americans suffering from MDD.
WHAT’S INSIDE
Here is an overview of the contents of this publication,
which has been prepared and reviewed by some of the na-
tion’s most respected experts in the field.
In “Prevalence and Economic Effects of Depression”
(page 9), Bernard Bloom, PhD, explores the economic
implications of MDD for individuals, health plans, em-
ployers, and society at large. Bloom presents direct medi-
cal costs and indirect costs of depression. Indirect costs
include not only disability and absenteeism, but also in-
teractive effects, costs with other diseases, and depres-
sion’s role in unemployment.
This is followed by an examination of current clinical
practice guidelines for treatment. Several national and in-
ternational organizations have published evidence-based
guidelines for the treatment of patients with MDD. In
“Clinical Practice Guidelines for Treating Depression in
Primary Care” (page 17), Christy Beaudin, PhD, and
Robert Burchuk, MD, review respected, commonly uti-
lized guidelines and discuss how they might be updated
to improve clinical outcomes given health plan environ-
ments and practice limitations. Guidelines, while key to
optimal treatment of patients with depression, are fre-
quently misunderstood as prescriptive rather than de-
scriptive, a phenomenon that contributes to their disuse.
Jeffrey Weilburg, MD, reviews the array of available
modern pharmacotherapies for depression — the selec-
tive serotonin reuptake inhibitors and serotonin norep-
inephrine reuptake inhibitors. In “An Overview of SSRI
and SNRI Therapies for Depression” (page 25), Weil-
burg examines their chemical structures, mechanisms of
action, contraindications, side effects, salient clinical trial
data, and indications for their use. Certain SSRIs have
been approved for social anxiety disorder, for instance.
Beginning on page 52, Ashok Raj, MD, reviews treatment
of this often comorbid condition.
We continue with a look at the pharmacoeconomics
of depression. Pharmacoeconomics is a tool with which
pharmacy and therapeutics committees, employers, and
payers can assess the value of pharmaceuticals. Pharma-
8 P&T DIGEST
coeconomics identifies, measures, and compares the costs
and outcomes of using a variety of pharmaceutical prod-
ucts as treatment alternatives for a particular disease
state. As Matthew Sarnes, PharmD, writes herein, more
and more decision makers in the managed care envi-
ronment rely on pharmacoeconomic analyses to make
informed formulary decisions. “Pharmacoeconomic
Evaluation of Antidepressant Therapies” (page 34) delves
into the specifics associated with the treatment of de-
pression, given specific assumptions and treatment
modalities.
As with many chronic diseases, the issue of patient
compliance with therapy is paramount, and depression
is no exception. “Adherence With Antidepressant Medi-
cation” (page 42), by Vincent Giannetti, PhD, provides an
in-depth look at reasons for patient noncompliance with
antidepressant medication regimens. The article goes be-
yond stating the problem, providing readers with data that
compare compliance rates for a number of SSRIs. Gian-
netti outlines some tested and reliable strategies to improve
compliance and, thus, clinical outcomes.
Finally, we look at how NCQA and the extent to which
HEDIS antidepressant medication management measures
improve the quality of health care for people with de-
pression (page 48). Benjamin Druss, MD, MPH, presents
national statistical averages and benchmark data. Druss
also explores roadblocks to improved HEDIS scores.
While depression may always be with us, there is much
that can and should be done to relieve the suffering of the
millions who will encounter this disease. Through phar-
macoeconomic models, we can start to appreciate the
economic ramifications of treatment; through the use
and careful interpretation of well-respected, nationally
published clinical practice guidelines, we can craft solid
treatment plans for depressed patients; with an under-
standing of the scope of the disease, diagnosed and undi-
agnosed, we can improve HEDIS performance measures
and, by extension, help patients with depression to reach
remission of symptoms and, hopefully, recover fully.
REFERENCES
CDC (U.S. Centers for Disease Control). Deaths: Final data for
2001. Natl Vital Stat Rep. 2001:52(3). Table 18.
Kessler RC, Berglund P, Demler O, et al. The epidemiology of
major depressive disorder: results from the National Co-
morbidity Survey Replication (NCS-R). JAMA. 2003;
289:3095–3105.
Lin EH, Von Korff M, Katon W, et al. The role of the primary
care physician in patients’ adherence to antidepressant
therapy. Med Care. 1995;33:67–74.
Olfson M, Marcus SC, Druss B, et al. National trends in the out-
patient treatment of depression. JAMA. 2002;287:203–209.
Olfson M, Marcus SC, Pincus HA, et al. Antidepressant prescrib-
ing practices of outpatient psychiatrists. Arch Gen Psychia-
try. 1998;55:310–316.
 E P I D E M I O LO G Y A N D E CO N O
Prevalence and Economic Effects
Of Depression
BERNARD S. BLOOM, PHD1
University of Pennsylvania
SUMMARY
The lifetime risk of major depression among
Americans is 17 percent, with as many as 10
percent suffering from depression in any 1-
year period.The author reviews the epidemi-
ology of depression, costs of treatment and
nontreatment, and its economic impact on
quality of life and daily function.This review
also examines ways to improve value for
money spent relative to this disease.
T
he World Health Organization (WHO) esti-
mates that about 340 million people worldwide
suffer from an episode of major depression each
year. It is the fourth leading cause of disability
worldwide. In the United States, about 18 million peo-
ple have had at least one episode of depression and sus-
tain a lifetime risk of about 17 percent (NIMH 2002).
Depression has a profound impact on population
health and health-system costs, individual and family
quality of life, activities of daily living, and daily func-
tioning, as well as on businesses, employers, and em-
ployees. About 15 percent of people with major depres-
sion commit suicide (Moller 2003).
Persons with depression tend to have multiple co-
morbidities, with substantial interactive effects relative to
suffering and cost. Because of the negative connotations
associated with mental health disorders, many depressed
1 Bernard S. Bloom, PhD, is research professor, Department
of Medicine, School of Medicine, University of Pennsylva-
nia. He also has appointments in the Health Care Systems
Department of the Wharton School and at the Leonard
Davis Institute of Health Economics, where he is a senior
fellow. Bloom’s research examines national and interna-
tional health policy issues, evaluating clinical, economic,
and quality-of-life outcomes with an emphasis on the el-
derly, physician decision making, the role of scientific medi-
cine in clinical practice, cost-effectiveness of care, and health
systems strategic planning, management and financing.
E PMI IDCESM I O LO G Y A N D E CO N O M I C S
persons shun treatment,
losing work days and per-
forming inadequately on
the job. These factors add
up to enormous human
and economic costs.
Depression is common,
persistent, often associated
with other chronic condi-
tions, underrecognized, and
undertreated. Conservative
estimates hold that clinical
depression (Kamlet 1995):
• Affects between 1 and Bernard S. Bloom, PhD
3 percent of the U.S.
population during any 6-month period
• Has a lifetime incidence of more than 15 percent
• Affects nearly twice as many females as males
• Recurs at least once among 50 percent of persons,
within 10 years of the initial episode
• Recurs a third time among 90 percent of persons
with 2 previous episodes
• Has a 40 percent relapse rate within 15 weeks among
persons with 3 or more lifetime episodes
• Has a relapse rate of 65 percent within the first year,
if untreated, among recurrent depressives
• Is the primary psychiatric disorder suffered by at
least 60 percent of suicide victims, accounting for
about 16,000 deaths annually
PREVALENCE
Pincus (2001) determined that the percentage of de-
pressed persons increases relative to the disease’s preva-
lence rate in the general population. Among general
medical patients, particularly hospital inpatients, the 1-
year prevalence rate of major depression is 2 to 3 times
that found in the community population.
Globally, in 1999, one person in six was expected to
suffer from major depression, with estimates that by
2020, major depression would be the second most im-
portant cause of disability worldwide (Davidson 1999).
DEPRESSION 9
 E P I D E M I O LO G Y A N D E CO N O M I C S
The National Comorbidity Sur-
vey (NCS) reported that nearly half FIGURE 1 Rate of diagnosis in primary care
of respondents had experienced at
least one psychiatric disorder 508 patients with depression studied
(Kessler 1994); most common was Self-reported score on Beck Depression Inventory
major depression, with a lifetime
prevalence of 17.3 percent and a 1- ≥9 (moderate to severe)
year prevalence of 10.3 percent.
The National Institute of Men-
tal Health (NIMH) Epidemiologic
Catchment Area (ECA) program
estimated lifetime prevalence of
major depression at 7 percent, and
reported that about one third of
major depressive patients sought
no treatment. In 1999, only about
10 percent received adequate
SOURCE: CALLAHAN 2002
doses of antidepressive drugs for a
sufficient period (Davidson 1999).
Depressed patients are 3 times
more likely to fail to adhere to treatment than non-
depressed patients, and most depressed persons suffer
chronic or recurrent episodes. About 40 percent of de-
pressed persons will meet diagnostic criteria for depres-
sion a year after the initial episode; 15 to 20 percent re-
main depressed for 2 years. About 60 percent of those
who experience an episode of depression will experi-
ence a second, and risk of recurrence increases with each
subsequent episode (Pincus 2001).
Quality of life and daily function
Even among people with minor depression, NCS re-
ported that symptoms interfere substantially with daily
activities in 18.1 to 52.3 percent of people, rising with in-
creased depression severity (Davidson 1999).Wells (1999)
found that depressed patients generally functioned at lower
physical and mental/emotional levels than those with 11
other common chronic conditions — asthma, diabetes, hy-
pertension, arthritis, migraines, as well as lung, neurolog-
ical, heart, gastrointestinal, vision, and back problems.
Those who self-reported 12-month depressive disorder
had lower mental and social function than those with any
of the other chronic conditions, and worse physical func-
tionality than patients with four of those conditions.
DIAGNOSIS AND TREATMENT
Typically in the United States, primary care physicians
diagnose and treat patients with depression and refer
them to specialists. The diagnosis of depression is asso-
ciated with higher primary care costs, but failure to di-
agnose it is associated with higher subsequent laboratory
and other diagnostic costs, as primary care physicians
strive to determine the cause of the patient’s symptoms.
10 P&T DIGEST
<9 (mild)
Rate of diagnosis 9.1%
27.6%
In the general medical setting, half of depressed pa-
tients are not diagnosed correctly (Croghan 1998). De-
pressives’ symptoms may be attributed to physical ill-
nesses; further, evidence suggests that physicians may
alter the stated diagnosis, resulting in differing claims,
medical records, and physician notes for the same patient.
Callahan (2002) found that primary care physicians
underdiagnose depression. As indicated in Figure 1, pri-
mary care physicians diagnosed depression in slightly
more than one fourth of patients who self-reported mod-
erate to severe depression, while making the diagnosis in
fewer than 1 in 10 patients whose depression was mild.
When patients’ own and primary care physicians’ diag-
noses failed to match, the physicians ordered laboratory
tests, increasing costs. The researchers concluded that pri-
mary care physicians were more likely to diagnose de-
pression accurately when the patient’s symptoms were se-
vere. Using a screening tool such as the Beck Depression
Inventory probably would increase depression diagnoses,
but it is unclear from this study what effect screening
would have on care processes, patient outcomes, or costs.
The NIMH Collaborative Depression Study reported
that only 34 percent with major depression lasting at
least 1 month received antidepressive treatment for at
least 4 weeks (Davidson 1999). In addition, the treatment
they received was adequate in only 23 percent of cases.
The investigators found a correlation between increased
depression severity and increased utilization of physician
services. Despite suffering recurrent episodes (mean 12.7
episodes per lifetime), only 35.3 percent of study pa-
tients sought professional help. The investigators observed
that depression symptoms can be controlled effectively,
but that the condition is often undiagnosed or inade-

quately treated in primary care
settings due to factors relating to FIGURE 2 Prevalence of depression in patients
the patient, provider, and system. with chronic disease
Depression also is treated in-
adequately by mental health spe-
cialists; only 45 to 49 percent of
Diabetes
patients diagnosed with depres-
sion receive adequate therapy (Da-
vidson 1999). The reasons found
for inadequate care were: Asthma
• Primary care physicians re-
ceive inadequate training in Stroke/myocardial
diagnosis and management
infarction
of depression; need to de-
velop the interpersonal skills
necessary to understand pa- SOURCE: PINCUS 2001
tients with depression; and
must develop the proper
mindset to treat depression as an illness.
• Patients with depression need counseling because
they lose initiative, often become passive, blame
themselves for their problems, and quickly become
noncompliant, all leading to adverse outcomes.
• Health care systems tend to perceive depression as
a short-term rather than chronic and recurrent ill-
ness, and most do not encourage monitoring pa-
tients throughout the treatment regimen. Instead,
they may discourage appropriate treatment and
follow-up care by reducing or denying coverage.
The major difficulties in treating depression (and any
mental illness) in primary care are rooted in organiza-
tional and financial arrangements. Behavioral health
carve-outs and risk-based physician payment often dis-
tort primary care physician decision making as a result
of the absence of real costs that physicians face when
making treatment decisions (Frank 2003).
COMORBIDITIES
Callahan (1997) suggested that patients with depres-
sion have a greater burden of comorbid nonpsychiatric
illness than people without depression. In addition, older
adults with depressive symptoms have nearly twice the
functional impairment as people without depression.
Figure 2 lists comorbidity rates for patients with com-
mon chronic conditions. Among diabetics, depression is
associated with poor glycemic control; increased risk of
complications, particularly heart disease; and higher over-
all health care costs. Asthmatics report decreased quality
of life. Among patients recovering from myocardial in-
farction, the mortality rate is 3 times higher in patients
with comorbid depression than for those without de-
E P I D E M I O LO G Y A N D E CO N O M I C S
27%
45%
40%
pression. Forty percent of high utilizers of health care have
been diagnosed with depression or dysthymia. These pa-
tients were found to represent the highest decile of con-
sumers of health care in a large HMO, responsible for 29
percent of primary care visits, 40 percent of inpatient
days, and 26 percent of prescriptions (Pincus 2001).
Economic consequences
A multinational study in Seattle; Barcelona, Spain;
Be’er Sheva, Israel; Melbourne, Australia; Porto Alegre,
Brazil; and St. Petersburg, Russia, demonstrated that the
economic consequences of depression are substantial in
the presence of medical comorbidity compared with de-
pression alone (Chisholm 2003). Researchers found that
comorbidity increased health care costs 17 to 46 percent
at five of six study sites. While they noted a strong asso-
ciation between depression and comorbidity, study lim-
itations did not allow them to establish a consistent trend
relative to the effect of psychiatric comorbidity on costs.
Croghan (1998) differentiated expenditures attribut-
able to depression from those related to comorbidities
that complicate depression. The investigative team re-
viewed patients’ records for comorbidities and total
health care charges for physician and other outpatient
visits, diagnostic tests, hospitalizations, and prescrip-
tions occurring within 1 year of the index antidepressant
prescription. In more than 95 percent of patients, they
found an indicator for at least one nondepression diag-
nosis. The most frequent of these, in 94.5 percent of
people, was nondepressive mental disorder; also com-
mon were ear, nose, mouth, throat, musculoskeletal sys-
tem, and skin disorders. Predicted expenditures for those
with at least one comorbid condition were significantly
higher than for patients without comorbidity.
DEPRESSION 11
 E P I D E M I O LO G Y A N D E CO N O M I C S
On average, treatment of depression alone represented
about 28 percent of total charges ($2,279 of $8,037) for
depressed patients with comorbidities. The most fre-
quently occurring comorbidity — mental illnesses other
than depression — accounted for approximately $1,600
of total charges. Other characteristics associated with
large incremental spending included neurological dis-
orders ($2,194), pregnancy and postpartum conditions
($1,697), and musculoskeletal conditions ($1,505). These
results indicate that while treating depression is expen-
sive, depression-related expenditures represent only
about one fourth of total cost. More recently, Kupfer
(2003) showed that comorbidities effectively double the
cost of caring for people with depression.
Over 4 years, Unützer (1997) studied people in an
HMO who were at least 65 years old. The investigators
found persistent depressive symptoms to be common,
and established an association between depressive symp-
toms and higher health care costs. Total health care costs
for seniors with significant depression symptoms were
about 50 percent higher than costs for those without de-
pression. They noted that primary care physicians may
be less likely to recognize depression in older patients or
may believe that they would not benefit from care.
Individuals with depression who are at least 60 years
old and who have a greater number of nonpsychiatric co-
morbid conditions visit ambulatory care centers and
emergency rooms more frequently, have more hospital-
izations, and have greater median total diagnostic test
charges for 1 year ($583 vs. $387) than those without de-
pression (Callahan 1997).
COSTS OF DEPRESSION
Using a human-capital approach, Greenberg (1993)
estimated the annual cost of de-
pression in the United States in
FIGURE 3 Annual costs associated with depression
1990 at $43.7 billion (Figure 3).
in the United States, 1990
Estimates were based on direct
costs of medical, psychiatric, and Direct
pharmacological care; mortality costs
costs associated with depression-
related suicide; and indirect costs
of morbidity, such as the costs as-
sociated with lost work days and
reduced productivity.
Researchers concluded that the $12.4
true economic burden might not
be realized, as many important
cost categories have yet to be esti-
mated. Future studies should ad-
dress comorbidity costs, reduced
quality of life, and patients’ and SOURCE: GREENBERG 1993
families’ out-of-pocket expenses.
12 P&T DIGEST
Davidson (1999) also estimated total cost of major de-
pression: nearly $44 billion in 1990. The investigators es-
timated that direct medical costs accounted for only
$12.4 billion (28 percent of total costs), for all inpatient
and outpatient services and pharmaceuticals. Early mor-
tality costs accounted for $7.5 billion. More than half of
the total cost, $24 billion, was attributed to absenteeism
and reduced productivity. This study did not account for
out-of-pocket expenses, excess hospitalizations, diag-
nostic tests, or costs associated with minor depression.
Absenteeism and reduced productivity
Depressive disorders pose a major occupational health
challenge, with implications for productivity, competi-
tiveness, absenteeism, insurance benefits utilization, and
medical care costs. Employers — the primary health care
payer in the United States — are concerned about in-
creased health-benefit costs. They often overlook the
benefits of treatment compared with continued illness
and inadequate diagnosis and treatment, however. Most
employers are unaware of how often depression con-
tributes to worker disability, the extent of its indirect
costs, and the availability of effective treatment options.
Further, because the workers’ compensation system and
courts have been slow to recognize depression as a work-
place disability, employers lack incentives to encourage
treatment and to implement preventive measures.
Goldberg (2001) concluded that depressed patients in
the United States were at least as functionally impaired
— in terms of physical, social, and role functioning —
as those with medical disorders such as hypertension, dia-
betes, advanced coronary artery disease, back problems,
angina, arthritis, breathing problems, and gastrointestinal
disorders. The investigators referred to a WHO study
Indirect costs
Mortality Morbidity
$23.8
$7.5

indicating that depressed patients
are 23 times more likely to have a FIGURE 4 Per-capita health care expenditures,
social disability 1 year after diag- patients with and without depression
nosis than are patients with physi-
10,000 covered employees and dependents, 1997
cal disability 1 year after diagnosis.
Conti (1995) found that depres-
sion was responsible for over half of
mental health diagnoses and claims
paid by a major Midwestern em-
ployer. Further, depression was re-
sponsible for more disability days
and 12-month recidivism than
were such physical complaints as
heart disease, diabetes, hyperten-
sion, and lower-back pain.
$2,059
In an international study, the
economic burden of untreated de-
pression also affected the work-
place (Chisholm 2003). When the
cost of lost work days is included, No depression claims
the economic burden increases SOURCE: GOLDBERG 2001
greatly. In the 3-months before
baseline assessment, study participants lost an average of
3.7 workdays (range 1.5–8.0 days), at an average cost of
$225 per participant. The authors acknowledged, but
did not include, the economic burden of partial workdays
that were lost.
In a study of 10,000 employees and dependents, Gold-
berg (2001) found mean annual health care expendi-
tures for patients with depression and related illnesses to
be more than 4 times that of patients with no depression
claims (Figure 4). The team noted that the accepted
prevalence rate of depression in the workplace — 2 per-
cent — is probably substantially underestimated.
Detected or hidden depression in the workplace leads
to reduced productivity, absenteeism, increased health
care resource utilization, job dissatisfaction, substance
abuse, and accidents. Kessler (1999) showed that de-
pressed workers lost 1.5 to 3.2 days more than other
workers in a 30-day period, at an average productivity
loss of between $182 and $395, based on an estimated
treatment cost of $402 per episode.
Employees with depression reported a mean of 5.6
hours per week of lost productive time, whereas those
without depression lost 1.5 hours. Extrapolation of these
data suggests that U.S. workers with depression cost em-
ployers an estimated $44 billion per year in lost produc-
tivity — $31 billion per year more than nondepressed
workers — and this estimate does not include labor costs
associated with short- or long-term disability.
A comparison of the effects of different common
chronic conditions on work impairment in the general
population showed that major depression was associated
E P I D E M I O LO G Y A N D E CO N O M I C S
$8,709
Added cost
of disability
leave due to
depression:
$5,000
Depression claims
with a higher likelihood of impairment than heart dis-
ease and hypertension (Kessler 2001). In addition, David-
son (1999) noted that 72 percent of all depressed persons
are in the workforce. The researchers estimate that annual
cost of depression per employee is nearly $4,900 and the
cost of appropriately treating persons with affective dis-
orders would offset direct costs by nearly $4 billion.
Work loss and disability
People with major depression have 4.78 times greater
risk of disability, while those suffering from minor de-
pression with mood disturbance (but not major depres-
sion) are at 1.55 times greater risk than those without de-
pression (Broadhead 1990). The investigative team
concluded that depression was an important predictor
for disability days during the year after diagnosis.
Druss (2000) studied the effects of heart disease, dia-
betes, hypertension, and back problems compared with
depression, and analyzed mental health costs, medical
costs, sick days, and total health and disability costs. The
per-capita cost of depression was significantly more than
that of hypertension or back problems, and comparable
to that of diabetes or heart disease (Figure 5, page 14).
The cost of treating employees with depression plus any
of the other conditions was 1.7 times higher than that of
treating employees with the same or other conditions
who were not depressed.
In an analysis of the NCS and the Midlife Develop-
ment in the United States Survey (MIDUS) data, Kessler
(1999) found that depression cost businesses between 1.5
and 3.2 more short-term disability days in a 30-day pe-
DEPRESSION 13
 E P I D E M I O LO G Y A N D E CO N O M I C S
riod than did workers without de-
pression. Major depression affects FIGURE 5 Comparison of health and disability costs
1.8 to 3.6 percent of the workforce
monthly. Salary-equivalent pro- Per capital annual costs ($)
ductivity loss averaged between Average annual sick days
$182 and $395. Using the esti-
mated cost of $402 for 30 days free
from depression provided by nor-
triptyline therapy, the research
4,388
team determined that effective
treatment offsets between 45 per- 3,732 5.4
cent ($182 of $402) and 98 percent
($395 of $402) of lost productiv-
ity costs. The team concluded that
because this study omitted indi-
rect costs, and because employers
bear the full brunt of lost produc-
tivity, effective treatment is cost-
5.4
effective for employers, especially
when other factors are considered Hypertension
— such as costs of short-term dis-
ability, early intervention, reduced SOURCE: DRUSS 2000
hospitalization, and other med-
ical care, and reduced absenteeism from work.
Rizzo (1996) estimated the economic effects of chronic
illness on productivity and the mitigating effects of pre-
scription drug therapy. The investigators estimated the
effects of medications on hourly wages and work days
lost, assuming average compliance. Effective drug ther-
apy achieved a net benefit, above the cost of therapy, for
employees with diabetes, heart disease, depression, and
hypertension (Figure 6). The investigators concluded
that benefits accrue because medications substantially
lower absenteeism among chronically ill workers.
Improving the mental health status of depressed medi-
cal patients who were high medical service utilizers alters
the course of functional disability (Von Korff 1992).
People with severe depression reported an average of
134 disability days per year at baseline; those with mod-
erate depression reported 77 days. For study participants
whose mental health improved by the 12-month follow-
up, average annual disability days fell 36 percent, from 79
to 51, for those with severe depression; for those with
moderate depression, disability days declined 72 per-
cent, from 62 to 18. Those with severe-unimproved de-
pression were more likely to have current major depres-
sion and to be unemployed.
HEALTH CARE RESOURCE UTILIZATION
Understanding health care resource utilization is vital
to understanding the total cost of depression. It starts with
investigating the cost-effectiveness of pharmacotherapy
but also addresses other factors affecting costs, such as the
14 P&T DIGEST
9.9
5,472 5,523
5,415 7.5
7.2 7.2
7.2 9.9 7.2 7.5
Back Depression Diabetes Heart
problems disease
value of psychotherapy, enhanced practice at the primary
care level, and improving adherence to a therapeutic reg-
imen. Borghi (2000) calculated that among people with
moderate or severe depression treated by a primary care
physician, overall cost of care for those who discontinued
depression treatment was about 50 percent higher than
for those who remained on their medications.
Henry (1997) noted that, in light of contemporary di-
agnosis and treatment abilities — and considering the
consequences of treatment failure, including attempted
and successful suicide — the indirect costs of depression
including mortality, may be 7 times as burdensome as the
direct medical costs. The authors advised that physicians
should focus primarily on clinical benefits when choos-
ing an antidepressant and offered that market price is only
one of many factors that influence medication choice.
They also warned that, in the long run, the trend to em-
phasize cost-containment first may increase costs.
Pharmacotherapy vs. psychotherapy
Kamlet (1995) investigated prospectively by random-
ized controlled trial the cost-utility of maintenance treat-
ment for depression. The researchers found imipramine
drug therapy to be more cost-effective than either inter-
personal psychotherapy (IPT-M) or placebo, improving
expected quality adjusted life years (QALYs) and reduc-
ing direct medical cost. Compared with placebo, neither
IPT-M nor combination therapy reduced direct medical
costs, but did improve QALYs.
Kamlet’s team determined a 0.0025 base case value to

the probability of suicide associated with a major de-
pressive episode. Using a $2.1 billion estimate of the di-
rect cost of depression in the United States in 1980 and
a 2 percent, 6-month prevalence rate, they estimated
$250 in direct per capita costs in 1980 dollars, or $500 in
1991 dollars, with a range of $200 to $1,500. To gener-
ate empirical results, the researchers used Monte Carlo
analysis for a baseline 40-year-old woman.
IPT-M and IPT-M+placebo increased expected QALYs
from 9.6 in the placebo group to more than 14, while drug
and combination therapy increased QALYs to between 15
and 22, depending on the magnitude of the estimated im-
pact of adverse side effects. IPT-M increased lifetime di-
rect costs by $13,000 (62 percent), which offsets the sav-
ings from a reduced number of future depressive episodes
and their associated costs. The cost of resulting health im- effective pharmacotherapy and psychiatric support, and
provements was less than $5,000 per QALY for both IPT-
M and imipramine, well below the $50,000 per QALY typ-
ically cited in the literature as the threshold to determine etal costs, particularly in terms of increased productiv-
if an intervention represents sufficient value for use of
health care resources.
Lave (1998) compared pharmacotherapy and psycho-
therapy in a randomized controlled prospective trial of
depressed persons. The researchers found that tricyclic
antidepressant therapy yielded slightly better quality-of-
life outcomes at a slightly lower cost compared with psy-
chotherapy. Both resulted in better outcomes than usual
care, at a higher cost and with no meaningful cost offsets. effective educational models for complex patients, and
Enhanced primary care
Pyne (2003) compared usual primary care for depres-
sion with enhanced intervention. Intervention enhance-
ment consisted of educating physicians, nurses, and of-
fice staff coordinators to assess, educate, and monitor
FIGURE 6 Net benefit in workplace of drug therapy
Drug-therapy expenditures more than offset lost productivity costs
Condition treated Annual net benefit to employer
Hypertension $286
Heart disease $633
Depression $822
Type 2 diabetes
SOURCE: RIZZO 1996
E P I D E M I O LO G Y A N D E CO N O M I C S
depressed patients during their acute depressive episode
and during the following year. The investigators assessed
the incremental effect of the intervention, defined as en-
hanced care minus usual care, on costs and QALYs. The
mean incremental cost-effectiveness ratio (CER) was
$15,463 per QALY. The team concluded that enhance-
ment was cost-effective compared with usual primary
care interventions and according to commonly accepted
cost-effectiveness thresholds.
IMPLICATIONS FOR PAYERS
Depression can be treated effectively, but providing ef-
fective therapy will likely increase up-front costs associ-
ated with educating primary care physicians and staff to
understand depression, screen effectively for it, provide
monitor adherence. Nevertheless, effective treatment,
according to available evidence, will reduce overall soci-
ity, reduced disability, fewer suicides, and better quality
of life for depressed persons and their families.
Callahan and colleagues added that primary care
physicians and mental health specialists should address
organic and psychiatric illnesses simultaneously to con-
trol health care costs and improve outcomes for depres-
sion patients (Callahan 1997). In particular, primary
care physicians need better treatment options and more
older adults need better access to mental health services.
MCOs could care for depressed patients more efficiently
by examining frequently occurring patient characteristics
that generally predict higher costs (Croghan 1998). Ap-
plying intensive services, in a collaborative model earlier in
the course of illness, would increase initial costs modestly
but also should improve outcomes,
therefore making a strong cost-
effectiveness argument.
Treating depressed persons
with appropriate care modalities
and regimens, including non-
pharmacological elements, is cost-
effective (Davidson 1999). Intro-
ducing a collaborative approach
to major depression would in-
crease response rates from 43.8
percent — with usual primary
care — to 74.4 percent, when a
primary care physician and psy-
chiatrist treat the patient together.
$1475
CONCLUSION
Depression is costly to treat ef-
fectively, but leaving it untreated is
DEPRESSION 15
 E P I D E M I O LO G Y A N D E CO N O M I C S
even more costly. Increased treatment nearly always in-
creases costs for every diagnosis. Most research on the
majority of diseases ultimately finds that early savings
evaporate over the long term when tested in the real
world. Improving patient health nearly always has up-
front costs. The conservative view would assume that im-
proved health would cost more but that the benefit to in-
dividuals and society may outweigh those concerns.
Thus, the issue is not whether treatment will improve
health outcomes and reduce expenditures, but whether
treatment and its benefits are worth the cost — the cost-
effectiveness question from the societal perspective.
Effective treatment starts with improving the mindset
and perceptions of primary care physicians regarding de-
pression, continues with effective treatment using anti-
depressants and psychiatric therapy in a coordinated ef-
fort, and leads to effective patient education, self-care,
and monitoring for adherence with established protocols.
Simply allowing the existing system of variable diag-
noses, care, and follow-up to continue will mean that
benefits are being forgone and resources wasted. The
problem is guaranteed to grow larger; already, in the
United States alone, between 1.8 and 3.6 percent of work-
ers suffer from major depression, between 17 and 21
percent of the workforce experiences short-term dis-
ability in any year, and between 37 and 48 percent suffer
short-term disability during their lifetimes.
If these estimates of disease burden are reasonably ac-
curate, direct and indirect expenditures for depression
will rise quickly and dramatically; low levels of benefit for
people with the disease, their families, employers, and so-
ciety will continue, however, if inadequate treatment
and follow-up by the health care system and patient ad-
herence to therapy are not corrected.
REFERENCES
Borghi J, Guest JF. Economic impact of using mirtazapine com-
pared to amitriptyline and fluoxetine in the treatment of
moderate and severe depression. Eur Psychiatry. 2000;15:
378–387.
Broadhead WE, Blazer DG, George LK, Tse CK. Depression, dis-
ability days, and days lost from work in a prospective epi-
demiologic survey. JAMA. 1990;264:2524–2528.
Callahan CM, Kesterson JG, Tierney WM. Association of symp-
toms of depression with diagnostic charges among older
adults. Ann Intern Med. 1997; 126:426–432.
Callahan EJ, Bertakis KD, Azari R, et al. Association of higher
costs with symptoms and diagnosis of depression. J Fam
Pract. 2002;51:540–544.
Chisholm D, Diehr PA, Knapp M, et al. Depression status, medi-
cal comorbidity and resource costs: evidence from an inter-
national study of major depression in primary care
(LIDO). Br J Psychiatry. 2003;183:121–131.
Conti DJ, Burton WN. The cost of depression in the workplace.
Behav Healthc Tomorrow. 1995;4:25–27.
Croghan TW, Obenchain RL, Crown WE. What does treatment
of depression really cost? Health Affairs. 1998;17:198–208.
16 P&T DIGEST
Davidson JRT, Meltzer-Brody SE. The underrecognition and
undertreatment of depression: what is the breadth and
depth of the problem? J Clin Psych. 1999;60(suppl 7):4–9.
Druss BG, Rosenheck RA, Sledge WH. Health and disability
costs of depressive illness in a major US corporation. Am J
Psychiatry. 2000;157:1274–1278.
Frank RG, Huskamp HA, Pincus HA. Aligning incentives in the
treatment of depression in primary care with evidence-
based practice. Psychiatr Serv. 2003;54:682–687.
Goldberg RJ, Steury S. Depression in the workplace: costs and
barriers to treatment. Psychiatr Serv. 2001;52:1639–1643.
Greenberg PE, Stiglin LE, Finkelstein SN, Berndt ER. The eco-
nomic burden of depression in 1990. J Clin Psych. 1993;54:
405–418.
Henry JA, Rivas CA. Constraints on antidepressant prescribing
and principles of cost-effective antidepressant use. Part 1:
depression and its treatment. Pharmacoeconomics. 1997;11:
419–443.
Kamlet MS, Paul N, Greenhouse J, et al. Cost utility analysis of
maintenance treatment for recurrent depression. Control
Clin Trials. 1995;16:17–40.
Kessler RC, Barber C, Birnbaum HG, et al. Depression in the
workplace: effects on short-term disability. Health Affairs.
1999;18:163–171.
Kessler RC, Greenberg PE, Mickelson KD, et al. The effects of
chronic medical conditions on work loss and work cutback.
J Occup Environ Med. 2001;43:218–225.
Kessler RC, McGonagle KA, Zhao S, et al. Lifetime and 12-
month prevalence of DSM-III psychiatric disorders in the
United States: results of the National Comorbidity Survey.
Arch Gen Psych. 1994;51:8–19.
Kupfer DJ, Frank E. Comorbidity in depression. Acta Psych
Scand. 2003;108(suppl 418):57–60.
Lave JR, Frank RG, Schulberg HC, Kamlet MS. Cost-effective-
ness of treatments for depression in primary care practice.
Arch Gen Psych. 1998;55:645–650.
Moller HJ. Suicide, suicidality and suicide prevention in affective
disorders. Acta Psychiatr Scand. 2003;418(suppl):73–80.
NIMH (National Institute of Mental Health). Depression. Avail-
able at: «http://www.nimh.nih.gov/publicat/
depressionmenu.cfm». NIH Publication No. 02-3561.
Reprinted 2002. Accessed March 31, 2004.
Pincus HA, Pettit AR. The societal costs of chronic major de-
pression. J Clin Psych. 2001;62(suppl 6):5–9.
Pyne JM, Rost, KM, Zhang M. et al. Cost-effectiveness of a pri-
mary care depression interview. J Gen Intern Med.
2003;18:432–441.
Revicki DA, Simon GE, Chan K, et al. Depression, health related
quality of life, and medical cost outcomes of receiving rec-
ommended levels of antidepressant treatment. J Fam Pract.
1998;47:446–452.
Rizzo JA, Abbott TA III, Pashko S. Labour productivity effects of
prescribed medicines for chronically ill workers. Health
Econ. 1996;5:249–265.
Unützer J, Patrick DL, Simon G, et al. Depressive symptoms and
the cost of health services in HMO patients aged 65 years
and older. JAMA. 1997; 277:1618–1623.
Von Korff M, Ormel J, Katon M, Lin EHB. Disability and depres-
sion among high utilizers of health care. Arch Gen Psych.
1992; 49:91–100.
Wells KB, Sherbourne CD. Functioning and utility for current
health of patients with depression or chronic medical con-
ditions in managed, primary care practices. Arch Gen Psych.
1999;56:897–904.
 D I AG N O S I S A N D T R E AT M E N T D I AG N O S I S A N D T R E AT M E N T

Clinical Practice Guidelines for

Treating Depression in Primary Care
CHRISTY L. BEAUDIN, PHD, LCSW, CPHQ1
ROBERT BURCHUK, MD2
PacifiCare Behavioral Health

SUMMARY prevention of relapse, and reduce treatment costs. It has

been demonstrated that patients have higher treatment
Clinical practice guidelines for depression rates for depression and bet-
offer the best available treatment options as ter outcomes when treated
in primary care practices
evidenced by randomized controlled clinical
that use evidence-based
studies. Although the use of evidence-based strategies (Wells 2000).
guidelines has been shown to improve treat- Clinical practice guide-
ment success, there are few data to suggest lines represent the best sci-
the extent to which they are used or their ef- ence underlying the art and
fect on treatment outcomes. Keeping guide- science of healing. As with
practice guidelines for any
lines current and increasing the probability
disease, those for depres-
of use in daily clinical practice are critical. sion are based upon clinical
evidence and refereed by

P
ractice guidelines for treating depression in pri- peers to ensure practical ap-
mary care have been available for a decade. Prac- plicability and consistency Christy L. Beaudin, PhD, LCSW,
titioners may think of them as prescriptive, before being released. The CPHQ
whereas their originators and organized medicine clinical evidence used in
intend them to offer guidance on clinical practice sup- guideline development is
ported by research and consistent with community stan- authoritative medical liter-
dards. Guidelines are not designed for dogmatic applica- ature with randomized
tion. However, using them can improve detection and controlled clinical studies
treatment of depression, reduce suicide risk, promote the as the standard.
Depression guidelines
1 Christy L. Beaudin, PhD, corporate director for quality im- for primary care practi-
provement at PacifiCare Behavioral Health, earned her doc- tioners are limited, with
torate in health services research from UCLA School of Pub- some simply listing med-
lic Health and her MSW from San Diego State University. ications without reference
Widely published, Beaudin also has worked as a consultant to psychotherapy for milder
and a psychiatric program administrator. She gave testimony cases. More widely available
to the Subcommittee on Evidence-Based Practices for the guidelines are those gener-
President’s New Freedom Commission on Mental Health. ated by professional soci- Robert Burchuk, MD
2 Robert Burchuk, MD, vice president and corporate medi- eties, organized delivery
cal director for PacifiCare Behavioral Health, is responsible systems, and health plans. Most depression guidelines
for medical management and quality improvement. He has used by clinicians and health care organizations are based
served as medical director for behavioral health programs solely or in part on the guideline issued by the American
at One Health Plan, Great West Life’s HMO, and Pruden- Psychiatric Association (APA 2000).
tial HealthCare Plan of California. He obtained his under- Health care organizations use practice guidelines to
graduate and medical degrees from Boston University. achieve improved treatment processes and outcomes.

DEPRESSION 17
 D I AG N O S I S A N D T R E AT M E N T
Their focus is on patient demographics, service needs,
and available resources, so self-generated guidelines may
differ from those issued by national medical societies. If a
health plan chooses to develop its own depression guide-
lines rather than adopt those of professional societies, the
guidelines should be based on a combination of the rec-
ommendations of professional societies, a review of the lit-
erature, scientific and expert presentations by clinicians
skilled in treating the disorder, and standards of care in the
community. Any organization embarking on this under-
taking should emphasize:
• Symptom assessment
• Suicide risk assessment and reassessment
• Medication options (if, when, how, how long to treat)
• Selection of modalities between drug options,
psychotherapy, and electroconvulsive therapy
• Monitoring adherence to medications and general
treatment plan
• Visit intervals
• Knowing when/how to switch or terminate treatment
• Clarifying maintenance treatment, if any
CONSIDERATIONS FOR DEPRESSION
GUIDELINES IN PRIMARY CARE
Fifteen years ago, depression was a diagnosis less com-
monly treated in primary care and more typically was re-
ferred to behavioral health specialists for treatment. With
the advent of selective serotonin reuptake inhibitors
(SSRIs) in the late 1980s, treatment of depression was
adopted more readily in the primary care setting — to the
point that the significant majority of antidepressant pre-
scriptions now are provided by nonpsychiatrists.
Studies on the inadequate treatment of depression
led the U.S. Department of Health and Human Services
to establish an Office of the Forum for Quality and Ef-
fectiveness in Health Care. As part of this effort, the
Agency for Health Care Policy and Research (AHCPR)
— today known as the Agency for Healthcare Research
and Quality (AHRQ) — issued guidelines for depression
in primary care in 1993 (these guidelines are no longer
considered current by AHRQ).
While AHRQ-sanctioned practice guidelines are avail-
able online through its National Guidelines Clearing-
house, there is currently no guideline available for the
treatment of depression. One AHRQ criterion for a guide-
line is that it must be no more than five years old. This sug-
gests that the APA depression guideline must be reviewed
by 2005. Maintaining or updating guidelines is expensive,
because organizations that issue them perform much of
the rigorous review of current research. They call on their
constituents to ensure that the guidelines are consistent
with current community standards of practice and known
developments in scientifically based research.
18 P&T DIGEST
Available guidelines
Depression guidelines have been disseminated by
health care organizations and professional associations.
Many have commonalities in assessment and treatment
strategies but diverge in the comprehensiveness of con-
sidering treatment strategies in a medical or primary
care setting where the treating clinician is not a behav-
ioral health specialist.
In 1993 and again in 2000, the APA issued its Practice
Guideline for the Treatment of Patients with Major De-
pressive Disorder. The APA guideline often is viewed as
the gold standard and is now updated on a fixed 5-year
cycle or, if needed, on the basis of important new clini-
cal information and research on the brain, human be-
havior, and therapies. The flow chart in Figure 1 depicts
APA recommendations for choosing treatment modali-
ties for major depressive disorder.3
The APA protocol suggests treatment options during
each phase of therapy, including starting and mainte-
nance doses of antidepressant medications. It also pro-
vides guidance for the clinician regarding when it is ap-
propriate to consider termination of therapy (APA 2000).
The most commonly distributed guidelines on de-
pression for primary care physicians are derived from or
rely heavily on the APA guideline, but are condensed
and may incorporate factors relevant to local delivery sys-
tem concerns or needs of special populations. Five guide-
lines presented below exemplify the different approaches
to guideline development: health care organization, com-
munity collaboration, statewide initiative, professional
association, and international development.
Health care organization
Brigham and Women’s Hospital
Depression: A Guide to Diagnosis and Treatment
Brigham and Women’s Hospital developed a depres-
sion guideline “to assist the primary care physician in
identifying patients in need of treatment and how to op-
timally treat them” (Brigham and Women’s Hospital
2004). The guideline offers recommendations for treat-
ing women with depression and is based on depression
research and the APA’s 2000 guideline. The Brigham and
Women’s guideline neither proscribes nor mandates treat-
ment options, but provides the primary care physician
with an algorithm for assessment and treatment planning
intended to meet the needs of the individual woman.
Within the guideline, the following are addressed:
• Medical impact of depression
• Medical conditions associated with depression
• Medications associated with depression
• Psychiatric disorders associated with depression
3 The entire APA guideline can be found online: «www.psych.org/
psych_pract/treatg/pg/Depression2e.book.cfm».
 D I AG N O S I S A N D T R E AT M E N T

• Life situations associated with depression
• Diagnostic strategy and treatment
• Commonly used antidepressants
• Depressive disorders unique to women
The Depression Management Algorithm offers the pri-
mary care physician a system driven by alarms related to
presentation of symptoms where treatment options in-
clude emergency treatment, a referral to psychiatry, a re-
ferral for substance abuse treatment, or an assessment for
medical comorbidities (Figure 2, page 20).
Community collaboration
Health Care Guideline:
Major Depression in Adults for Mental Health Providers
In Minnesota, where large group practices are com-
mon, 30 medical groups adopted a single set of guidelines.
One objective of the Institute for Clinical Systems Im-
provement (ICSI), a collaboration of health care organi-
zations, is to accelerate implementation of best clinical
practices. An ICSI effort is underway to improve depres-
sion treatment. Its goals include improving diagnosis and
assessment by using rating scales, such as the Hamilton
Rating Scale for Depression, and increasing adherence to
therapy and maintenance of treatment via follow-up con-
tacts by health care providers and their staffs (ICSI 2003).
The ICSI guideline is summarized in Table 1, page 21.
Statewide guideline
Colorado Major Depression Disorder in Adults
Diagnosis and Treatment Guidelines
Universal adoption of a statewide guideline may pro-
vide advantages, such as a single point of reference for all
clinicians and creation of a natural environment in which
to conduct impact research. In Colorado, health care

FIGURE 1 APA recommendations for choice of treatment modality

#1 Should a specific effective psychotherapy be provided?

Do not include a Include specific
specific effective No • Mild to moderate depression: preferred as solo treatment Yes effective
psychotherapy in or in combination psychotherapy in
treatment plan • Moderate to severe depression: in combination with medication or treatment plan
and go to question 2 electroconvulsive therapy (ECT) IF psychosocial issues are important and go to question 2
and/or IF preferred

#2 Should medication be provided? Include

Do not include medication in
No • Mild depression: IF preferred as solo treatment Yes treatment plan
medication in
treatment plan • Moderate to severe depression: with or without a specific effective and continue to
and go to psychotherapy unless ECT is planned other treatment
question 3 considerations (not
• Psychotic depression: combination of antipsychotic medication shown in this figure)
and antidepressant medication, or ECT

#3 Should medication and a specific effective psychotherapy both be

provided?
Do not include Include
• Mild depression: medication and a
medication and a • IF preferred as combination treatment
specific effective specific effective
No • History of partial response to single modality Yes psychotherapy in
psychotherapy in • Poor compliance
treatment plan treatment plan
and go to • Moderate to severe depression: and continue to
question 4 • Prominent psychosocial issues other treatment
• Interpersonal problems considerations
• Personality disorder
• Poor compliance

#4 Should ECT be provided?

Do not include ECT • Chronic, moderate to severe depression: with or without a specific Include ECT
in treatment plan No effective psychotherapy, IF patient prefers. Yes in treatment plan
and continue to • Severe depression and any of the following: and continue to
other treatment • Psychotic features other treatment
considerations • Patient prefers considerations
• Previous preferential response, need of rapid antidepressant
response, intolerance of medication

Go to other treatment considerations

SOURCE: APA 2000

DEPRESSION 19
 D I AG N O S I S A N D T R E AT M E N T

providers and plans agreed 3 years ago on a single set of
criteria to guide the treatment of depression. The Col-
orado Clinical Guidelines Collaborative developed its
Major Depression in Adults Diagnosis and Treatment
Guidelines. Notably, the brevity of the Colorado guide-
lines addresses the low success rate in diagnosis and ini-
tial treatment of depression. The short-form guidelines
(Figure 3, page 22) provide a thumbnail description of
an appropriate treatment plan by graphically depicting
a timeline with intervals for follow-up visits; a physician
can use this tool to direct the future schedule for patient
appointments during different treatment phases —
acute, continuation, and maintenance, if indicated. It
also includes a classification scheme for antidepressants,
depicting dosages, potential adverse effects, and contra-
indications for certain medications.
Professional association
American College of Physicians (ACP)
Pharmacological Treatment of Acute Major Depression
and Dysthymia
The ACP develops its evidence-based practice guidelines
using a rigorous development process. The ACP guidelines
are produced through scientific policy staff, a steering
committee (Clinical Efficacy Assessment Subcommittee),
and in collaboration with scientists. ACP released two
publications to support its depression guidelines:
• Clinical Guideline Part 1: Pharmacological Treat-
ment of Acute Major Depression and Dysthymia
(Snow 2000)
• Clinical Guideline Part 2: A Systematic Review of
Newer Pharmacotherapies for Depression in Adults:
Evidence Report Summary (Williams 2000)
International development
Clinical Practice Guidelines for the Treatment of
Depressive Disorders
The Canadian Network for Mood and Anxiety Treat-
ments, in collaboration with the Canadian Psychiatric As-
sociation, developed a guideline that outlines strategies for
treatment of depressive disorders. Published as a supple-
ment to the Canadian Journal of Psychiatry, the guideline
was a national initiative and a collaborative effort among
providers with input from international experts (CPA/
CANMAT 2001). The guideline offers a description of
prevalence and health burden, principles of management,
psychotherapy, comorbidities, medication, and biologic
treatments. Notably, the guideline reports on the needs of
special populations for which there may be fewer evidence-
based treatment recommendations in the literature.
USE/DISUSE OF GUIDELINES
Barriers to using guidelines are numerous, and range
from the philosophical to the most mundane practical
considerations — including how to locate the guidelines
when they are needed (Table 2, page 23). Most medical

FIGURE 2 Depression management algorithm

Brigham and Women’s Physician Hospital Organization

Major risk factors for recurrent depression
The presence of any of the following should direct
SIGE CAPS:
the primary care physician to consider, strongly, main-
tenance use of antidepressants and/or consider a
If yes, No alarms
consultation with a psychiatrist:
consider
presence of 1. Family history of recurrent depression
alarms: 2. Family history of bipolar disorder
3. Personal history of recurrence within 1 year after
Consider psychosis, mania, discontinuing effective treatment
Alarms are eating disorder, obsessive
compulsive disorder, as well 4. Onset of major depressive episode before age 20
present
as history of hypomania 5. Severe, sudden, or life-threatening depressive
episode (e.g., suicide attempt)

Send to emergency
department If yes, refer If no, consider alcohol
to psychiatry or substance abuse

If no, consider
If yes, refer to substance medical etiologies
abuse center

If yes, treat If no, begin antidepressant

as indicated treatment
SOURCE: BRIGHAM AND WOMEN’S HOSPITAL 2004

20 P&T DIGEST

TABLE 1 ICSI Guideline for Depression, Major, in Adults in Primary Care
Scope and target population: adults >18 years
Clinical highlights for individual clinicians
1. A reasonable way to evaluate whether a system is suc-
cessfully functioning in its diagnosis, treatment plan,
and follow-up of major depression is to consider:
• How well the diagnosis is documented
• How well the treatment team engages and edu-
cates patients/families
• Whether the system documents ongoing patient
contact
• Whether the system measures/documents outcomes
2. Presentations for major depression include:
• Multiple somatic complaints, weight gain/loss, mild
dementia
• Multiple (>5/year) medical visits; more than one
organ system affected with absence of physical
findings
• Fatigue
• Work or relationship dysfunction/changes in inter-
personal relationships
• Sleep disturbances
3. Effective treatment for some patients presenting with
a major depressive disorder may differ significantly
from other depressed patients.When assessing a pa-
tient, consider asking about manic or hypomanic
episodes.
4. If the patient is not experiencing a significant reduc-
tion of symptoms after 4–6 weeks of treatment, other
treatment strategies should be considered.
5. The key objectives of treatment are:
• To achieve remission of symptoms in the acute
treatment phase for major depression
• To reduce patient relapse and reduction of symp-
toms
• To return to previous level of occupational and
psychosocial function
• Antidepressant medications and/or referral for
psychotherapy are recommended as treatment for
major depression without coexisting medical condi-
tions, substance abuse, or other specific psychiatric
comorbidities. Physical activity and tailored patient
education are also useful tools in easing symptoms
of major depression.
SOURCE: ADAPTED FROM ICSI 2003
practitioners still use paper — not digital — records, and
with the literature they see daily, paper proliferates.
Those physicians who perceive guidelines as infring-
ing on their autonomy may negatively influence the use
of guidelines. Some remain skeptical about outcomes
research and evidence-based medicine in general, espe-
cially more seasoned clinicians; these practitioners are
often unwilling to abandon practice traditions and office
D I AG N O S I S A N D T R E AT M E N T
6. When antidepressant therapy is prescribed, medi-
cation adherence and completion is critical.The pa-
tient should be advised of the following:
• Most people need to be on medication at least 6
months.
• It may take 2–6 weeks before improvement is
seen.
• Take the medication as prescribed, even after the
patient starts feeling better.
• Do not stop taking the medication without call-
ing your provider. Side effects can be managed by
changes in the dose or dosage schedule.
Priority aims and suggested measures for health care
systems
1. Increase the use of DSM-IV TR criteria in the detection
and diagnosis of major depression in primary care.
2. Increase the functional status of patients with major
depression (e.g., Hamilton [HAM-D] scores).
3. Increase the percentage of patients with major de-
pression who stay on antidepressants for clinically ap-
propriate periods.
4. Improve the adherence and maintenance of appropri-
ate treatments for patients diagnosed with major de-
pression by having follow-up contacts with a health
care professional.
5. Increase the assessment for major depression of pri-
mary care patients with more than 5 visits in the past
year with problems in more than one organ system.
Additional background
Depression is a treatable cause of pain, suffering, disabil-
ity, and death, yet primary care providers detect depres-
sion in only one third to one half of their patients with
major depression. Anxiety disorders (panic disorder and
generalized anxiety disorder) are associated with dimin-
ished well being, increased substance abuse, increased
medical care utilization, and suicide attempts at rates
often exceeding other psychiatric problems including
major depression.This guideline stresses early suspicion
of the two-question screen, depression, and a “positive
diagnosis,” by asking simple key interview questions.
culture (Diamond 1998). But even if a practitioner wishes
to use guidelines, there may be issues surrounding patient
care (e.g., setting, funding) that make it impossible to im-
plement the guideline’s recommendations (TAC 2001).
The reason for issuing the guidelines must be explained
to physicians, along with the expected benefits of using
them. A clear rationale for the guidelines helps dispel
fears about lack of autonomy. Further, new guidelines
DEPRESSION 21
 D I AG N O S I S A N D T R E AT M E N T
FIGURE 3 Colorado Clinical Guidelines Collaborative protocol for major depression in adults
Common Symptoms
• Pains and aches
• Low energy
• Apathy, irritability, anxiety, sadness
• Sexual complaints
• Disrupted sleep patterns routine medical screens
• Vague GI symptoms
Depression Criteria (DSM IV):
5 or more in same 2 weeks, including at
least one of the first two symptoms:
• Depressed mood
• Marked diminished interest/pleasure
• Significant weight gain or loss
• Insomnia or hypersomnia
• Psychomotor agitation or retardation
• Fatigue or loss of energy
• Feelings of worthlessness or inappropriate guilt
• Diminished concentration or indecisiveness
• Suicidal ideation (thoughts, plans, means, intent)
If imminently suicidal, consider psych consult,
emergency hold, 911, and/or
psychiatric inpatient evaluation.
Consider Comorbid Medical Conditions
& Other Psychiatric Disorders
Carefully screen for Bipolar and Substance Abuse
Treatment and/or Referral Options:
• Medications -- especially for moderate to severe
and/or chronic symptoms
• Referral to Outpatient Psychotherapy-- suitable
for mild to moderate symptoms.
• Combined medication and psychotherapy -- for Determine method of
more severe symptoms and incomplete response to
treatment:
either medications or therapy.
Medication Selection &
Dosage Considerations: Educate patient about:
• Existing medical & psychiatric conditions (see • medication side effects
Appendix 1 ) • importance of
• Side effects
• Lethality for suicidal patients • not character defect/
SOURCE: COLORADO CLINICAL GUIDELINES COLLABORATIVE 2003
should be geared toward specific populations and health
care settings. There are important differences among
guidelines and approaches for different age groups. Well-
designed guidelines offer both an in-depth version for ed-
ucational purposes for the clinician and a shorter, quick-
reference version for use in the office. Clinicians do not
have time to pore over lengthy guidelines.
Overproliferation of guidelines is a problem. While a
practitioner in an integrated health system probably has
two sets of guidelines, a primary care physician in a group
practice may be expected to use different sets of guide-
lines from all the health plans with which he or she con-
tracts. Sensory and information overload may become
analogous to too many brands on a shelf, which makes
intelligent product selection difficult.
Guidelines can be integrated into quality improvement
(QI) programs and activities. In fact, the National Com-
mittee for Quality Assurance requires that MCOs adopt
behavioral health guidelines and has an established frame-
work for integrating practice guidelines into QI manage-
ment and improvement (see Table 3). Recent research
suggests that QI programs targeting depression can make
a difference in outcomes (Sherman 2004, Wells 2004).
22 P&T DIGEST
Attend to common
symptoms of • Monitor treatment
depression during
response using criteria
and/or depression scale
• Ongoing patient education Acute Treatment Phase (Wk 1 - Wk 12)
on course of illness • symptom reduction expected by Wk 6 to 12
and compliance • first follow-up appt after evaluation in Wk 1- 3
• next fo llow-up appts/ contacts every 2 - 4 Wk
• evaluate response by Wk 6
Augment or change Partial or no
Complete symptom resolution
Improvement
treatment
• increase dosage
• try different medication Continuation Phase (Mo 4 -Mo 9)
• refer for therapy
Confirm Diagnosis • obtain consult • begins after symptom resolution observed
using criteria and/or • continue medications at full strength
depression scale Complete symptom • schedule appt/ contact every 2-3 Mo
( see Appendix 2 ) Partial or no resolution
Improvement
Discontinue with taper or
proceed with maintenance
Obtain psych consult or Maintenance Phase (Mo 9 and on)
refer to mental health • at-risk for relapse based on history or genetic
specialty care disposition
• aimed at preventing relapse
• continue medications for one to several years
Psych Consult/Referral Considerations
• Guideline not suitable for patient
• Failed or only partial response to one or more medication trials
• Co-administering second anti-depressant
• Active suicidal, homicidal, self injurious behavior
• medication
• Psychotic or bipolar depression
• psychotherapy
• Severe depressive symptoms
• both
• Complex psychological issues
• Second opinion desired
• Co-existing substance abuse
• Medically unstable geriatric patient
compliance
personal weakness Final Short Form
Guideline development and implementation can be a
significant component of a QI program where an orga-
nization targets prevalent or higher-risk diagnoses or dis-
orders with the goal of improving treatment process or
outcomes. As an example, the Minnesota effort marries
guidelines with QI and focuses on building a relationship
with providers to deploy guidelines and produce im-
proved process and/or outcomes.
OPPORTUNITIES AND CONSIDERATIONS
Despite the development of the second-generation
SSRIs and serotonin norepinephrine reuptake inhibitors,
as well as such nonmedical therapies as problem-solving
psychotherapy, treatment practices still lag new findings
from clinical studies. A number of studies document in-
adequate care of depression (NIMH 1999).
National initiatives in managed care have been un-
dertaken to improve depression treatment. AHRQ,
through its Patient Outcomes Research Team on de-
pression, found that only one fourth of patients with de-
pression received appropriate care in a primary care set-
ting. This was attributed in part to the fact that primary
care patients often resist psychiatric labeling; further,
 D I AG N O S I S A N D T R E AT M E N T

primary care physicians may lack depth of training in

psychiatric disorders and may face barriers to referring
their severely depressed patients to specialty mental
health services (AHRQ 2003).
Can the lack of effective depression care in the United
States be remedied by guideline adherence? This is un-
known; further research is necessary to determine the ex-
tent to which treatment quality initiatives and treatment
guidelines improve patient care in any setting. It is still
unknown whether the pioneering depression treatment
guidelines of the 1990s have influenced depression treat-
ment by the average primary care physician. Several areas
of opportunity exist for improving care for patients with
depression. These address systemic issues and practical
considerations, in addition to clinical best practices.
Systems issues. In 2001, President Bush established the
President’s New Freedom Commission on Mental Health
to provide a road map for improving mental health ser-
vices within the budgets of existing funding streams.
Two years later, the commission reported that the treat-
ment of mental health in the United States was in alarm-
ing disarray, partially because of the fragmentation of the
mental health service delivery system. The report stated
that many outdated and ineffective treatments were still
being used, and that there was a lag between state-of-the-
art treatment options and actual clinical practice.
TABLE 2 Barriers to guideline
implementation
Practice level
• Primary care physician time pressures
• Difficulty locating the guidelines in the office
• Lack of time/resources for implementing the
practice recommendations
• Patient comorbidities that make generalized
guidelines risky to follow
System level
• Insurance coverage not available for recom-
mended therapy
• Organizational fragmentation in the health care
system
• Guidelines not well-integrated into the health
care system
• Lack of timely dissemination of guidelines
to practitioners
One New Freedom recommendation notes: “The
workforce will be trained to use the most advanced tools
for diagnosis and treatments… translating research into
practice.” This would be accomplished by circulating
evidence-based protocols beyond the mental health sub-
specialties into the general health care setting and into

TABLE 3 NCQA framework for quality improvement

As part of its quality improvement program, an organization utilizes clinical practice guidelines to help practitioners
and members make decisions about appropriate health care for specific clinical circumstances. Key elements of prac-
tice guidelines and quality improvement include:

Element
The organization adopts clinical
practice guidelines for acute,
chronic, and behavioral health
care, relevant to its population.
The organization establishes a
clinical basis for its guidelines
by:
• Using evidence-based clinical
practice guidelines
• Using an appropriate body
for approval
The organization annually mea-
sures performance against at
least three clinical practice
guidelines, one of which relates
to behavioral health.
Explanation
Guidelines that the organization adopts and promotes to practitioners to improve
health care and reduce unnecessary variations in care.There must be evidence
that the organization has adopted at least three clinical practice guidelines for the
treatment of acute, chronic, and behavioral health conditions.
Evidence-based guidelines are clinical practice guidelines known to be effective in
improving health outcomes. Effectiveness of guidelines is determined by scientific
evidence or, in the absence of scientific evidence, professional standards or, in the
absence of professional standards, expert opinion.The organization must adopt
guidelines from recognized sources or involve board-certified practitioners from
appropriate specialties in the development or adoption of its own clinical practice
guidelines that are not from recognized sources.
Routinely measuring specific aspects of care addressed in the clinical practice
guidelines determines whether the organization and its practitioners follow the
guidelines. It is preferable to measure more aspects of performance. If the organi-
zation chooses to use only two measures to assess performance, the measures
must relate to important aspects of the guideline most likely to affect care. Be-
cause clinical practice guidelines describe the process of care, the intent of this
standard is to measure at least two important clinical process elements that relate
directly to the clinical guideline.

DEPRESSION 23
 D I AG N O S I S A N D T R E AT M E N T
TABLE 4 Components of effective practice guidelines
• Define the clinical setting for which the guidelines are appropriate
• Consider all important treatment options and their consequences
• Identify the evidence underlying the guidelines
• Provide references with publication dates
• Describe explicitly how the guidelines were derived
• Clearly recommend interventions
• Allow review by intended users as well as peers
• Devise a mechanism for dissemination, evaluation, and updates
MODIFIED FROM SCALZITTI 2001
schools. To accomplish this, the commission encour-
aged technology-based educational and information ac-
cess (New Freedom 2003).
Increasing the recovery rate. This is a foremost oppor-
tunity for improvement. Guidelines can create such
opportunities by improving symptom recognition, in-
cluding proactive screening questions, providing for co-
ordination between behavioral health and medical prac-
titioners, and reducing the complexity of managing other
conditions and illnesses that accompany depression.
Flexibility. Good guidelines take into account the wishes
of the patient for treatment choices and cover all forms of
therapy, including pharmacotherapy and psychotherapy.
Table 4 lists characteristics of effective guidelines.
The process of developing guidelines involves many
steps, checks, and balances. There is typically a 3 to 5 year
lag between guidelines revisions by professional associ-
ations. For guidelines to be effective, there must be a feed-
back mechanism in place so “practice-based evidence”
of treatment efficacy can be documented and incorpo-
rated into updated guidelines as new evidence.
CONCLUSION
Clinical practice guidelines that are evidence-based,
rather than anecdotal or consensus-based, can advance
success in the diagnosis and treatment of depression in
the primary care setting. Several well-documented and
carefully developed guidelines exist; these documents
serve to assist clinicians with the art and science of heal-
ing. A guideline’s recommendations should be consid-
ered in conjunction with a patient’s progress in recovery.
Treatment guidelines are not mandates. In and of them-
selves, they cannot cure depression or prevent recurrences.
Psychiatric disorders such as depression can be, by nature,
chronic diseases. As with many other diseases addressed
by the primary care practitioner, early detection, inter-
vention, and treatment prevent relapse and hospitaliza-
tion, and reduce emotional and financial burdens to the
individual and society. As guidelines are updated, con-
sideration should be given to both treatment advances and
practical considerations that will encourage their use.
24 P&T DIGEST
REFERENCES
AHCPR (Agency for Health Care Policy and Re-
search) Depression Guideline Panel. Depression
in Primary Care: Volumes 1 and 2 (Clinical
Practice Guideline No. 5). Rockville, Md.:
AHCPR. 1993. Available at: «http//www.
mentalhealth.com». Accessed March 10, 2004.
AHRQ (Agency for Healthcare Research and Qual-
ity). Researchers examine efforts to improve
primary care for depression. Available at:
«http://www.ahrq.gov/research/mar03/
0303ra16.htm». Accessed March 10, 2004.
APA (American Psychiatric Association). Practice
guideline for the treatment of patients with
major depressive disorder (revision). Am J
Psychiatry. 2000;157(4 suppl):1–45.
Brigham and Women’s Hospital. Depression: A Guide to Diagnosis
and Treatment. Clinical guideline. 2004. Available at:
«http://www.brighamandwomens.org/medical/
handbookarticles/depression/depression_frame.asp». Accessed
April 2, 2004.
CPA (Canadian Psychiatric Association) and CANMAT (Canadian
Network for Mood and Anxiety Treatments). Clinical Practice
Guidelines for the Treatment of Depressive Disorders. 2001.
Available at: «http://www.cpa-apc.org/Publications/Clinical_
Guidelines/depression/clinicalGuidelinesDepression.asp».
Accessed May 27, 2004.
Colorado Clinical Guidelines Collaborative. Major Depression Dis-
order in Adults. Clinical guideline. 2003. Available at:
«http://www.coloradoguidelines.org/pdf_files/DepShort03.
pdf». Accessed March 10, 2004.
Diamond F. You can drag physicians to guidelines but you can’t
make them comply (mostly). Manag Care. 1998;7(9):14–22.
ICSI (Institute for Clinical Systems Improvement). Major Depres-
sion in Adults for Mental Health Care Providers. Clinical
guideline. Bloomington, Minn.: ICSI. 2003. Available at:
«http://www.icsi.org/knowledge/detail.asp?catID=29&itemID=
179». Accessed March 10, 2004.
NCQA (National Committee for Quality Assurance). Standards and
Guidelines for Accreditation of MCOs. Washington: NCQA. 2002.
New Freedom (President’s New Freedom Commission on Mental
Health). Report, 2003. Available at: «http://www.
mentalhealthcommission.gov». Accessed March 9, 2004.
NIMH (National Institute of Mental Health). Office of the Surgeon
General, Center for Mental Health Services. Mental Health: A
Report of the Surgeon General. 1999. Washington: NIMH.
Scalzitti DA. Evidence-based guidelines: application to clinical prac-
tice. Phys Ther. 2001;81:1622–1628.
Sherman SE, Chapman A, Garcia D, Braslow JT. Improving recogni-
tion of depression in primary care: a study of evidence-based
quality improvement. Jt Comm J Qual Saf. 2004 Feb;30:80–88.
Snow V, Lascher S, Mottur-Pilson C. Pharmacologic treatment of
acute major depression and dysthymia clinical guideline, Part
1. Ann Intern Med. 2000;132:738–742.
TAC (Technical Assistance Collaborative). Turning Knowledge Into
Practice. 2001. Boston: TAC and the American College of Men-
tal Health Administration. Available at: «http://www.tacinc.org/
index/viewPage.cfm?pageId=114». Accessed March 10, 2004.
Wells KB, Sherbourne C, Schoenbaum M, et al. Impact of dissemi-
nating quality improvement programs for depression in man-
aged primary care: a randomized controlled trial. JAMA.2000;
283:212–220.
Wells K, Sherbourne C, Schoenbaum M, et al. Five-year impact of
quality improvement for depression: results of a group-level
randomized controlled trial. Arch Gen Psychiatry. 2004;61:
378–386.
Williams JW, Mulrow CD, Chiquette E, et al. A systematic review of
newer pharmacotherapies for depression in adults: evidence re-
port summary clinical guideline, Part 2. Ann Intern Med. 2000;
132(9):743–756.
 P H A R M ACOT H E R A P Y
An Overview of SSRI and SNRI
Therapies for Depression
JEFFREY B. WEILBURG, MD1
Massachusetts General Hospital
SUMMARY
Selective serotonin reuptake inhibitors
(SSRIs) and serotonin norepinephrine re-
uptake inhibitors (SNRIs) are used widely to
treat mood and anxiety disorders. Indica-
tions, pharmacologic characteristics, and
dosing and administration are outlined. Be-
cause more patients receive SSRIs in general
medical versus psychiatric settings, this chap-
ter includes information relevant to both.
D
epressive episodes may be precipitated by
stressful events and relieved by psychotherapy.
Yet major depressive disorder (MDD) and re-
lated mood disorders are serious medical ill-
nesses that have a significant negative impact on mor-
bidity, mortality, role function, and quality of life (Druss
2000). Even mild depressive symptoms can affect em-
ployee productivity and health care costs significantly.
MDD has a genetic basis and presents with remarkable
similarity and frequency across cultures. Criteria for di-
agnosing MDD are provided in Figure 1 on page 26.
In addition to MDD, mood disorders outlined in the
Diagnostic and Statistical Manual of Mental Disorders,
Fourth Edition, Text Revision (DSM-IV-TR) (APA 2000)
include dysthymic and bipolar disorders, and mood dis-
orders caused by a general medical condition or sub-
1Jeffrey B. Weilburg, MD, is assistant professor of psychiatry
at Harvard Medical School, a staff psychiatrist at Massa-
chusetts General Hospital, and associate medical director of
Mass General Physicians Organization. The author of more
than 50 articles, chapters, and posters, Weilburg’s research in-
terests include the economic consequences of inadequate
treatment of depression, the relationship between treatment
adequacy and the HEDIS antidepressant-use method, and
an ongoing project providing feedback to primary care physi-
cians about adequacy-of-treatment rates. Weilburg received
his MD from George Washington University. He is board cer-
tified by the American Board of Psychiatry and Neurology.
P H A R M ACOT H E R A P Y
stance abuse. In DSM-IV-
TR, premenstrual dysphoric
disorder (PMDD) was listed
among depressive disorders
not otherwise specified, but
a consensus panel has de-
termined that PMDD is a
distinct clinical entity. The
U.S. Food and Drug Ad-
ministration subsequently
embraced the concept, and
agents have been marketed
for the treatment of PMDD.
Dysthymia, sometimes
called “minor depression,” Jeffrey B.Weilburg, MD
is a chronic mild to moder-
ate form of depression that persists for at least 2 years (1
year in adolescents and children). Those patients with
dysthymia are at risk of developing MDD; treatment for
dysthymia is nearly identical to that for MDD.
Selective serotonin reuptake inhibitors (SSRIs) and
serotonin norepinephrine reuptake inhibitors (SNRIs)
are important in the treatment of mood and anxiety dis-
orders (Table 1 on page 27 lists approved indications).
These disorders often coexist and thus may be difficult for
clinicians in a general medical setting to distinguish from
depression. The DSM-IV-TR list of anxiety disorders in-
cludes acute stress disorder; agoraphobia without history
of panic disorder; anxiety disorder related to a medical
condition or substance abuse; generalized anxiety disor-
der (GAD); obsessive-compulsive disorder (OCD); panic
disorder, with or without agoraphobia; posttraumatic
stress disorder (PTSD); specific phobia (e.g., acrophobia,
arachnophobia); and social anxiety disorder (SAD).
The goal of treatment is symptom remission or full re-
sponse to medication, defined as a complete or nearly
complete resolution of symptoms marked by a return to
premorbid baseline social and occupational function
and a sustained period of symptom relief. Guidelines
for treating anxiety disorders and mood disorders in-
clude appropriate use of antidepressants to achieve and
maintain remission, which typically involves adequate
DEPRESSION 25
 P H A R M ACOT H E R A P Y

FIGURE 1 Algorithm for diagnosis of major depressive disorder (MDD)

During the past month, have you During the past month, have you
often been bothered by feeling often been bothered by having little
No
down, depressed, or hopeless? interest or pleasure in doing things?

Patient is unlikely
Yes No to have MDD

A diagnosis of MDD requires ≥5 of the following symptoms, including depressed mood or anhedonia, during the same
2-week period, causing clinically significant distress or impairment in social, occupational, or other areas of functioning.

Symptom Diagnostic criteria

Depressed mood Depressed mood most of day, nearly every day
Anhedonia Markedly diminished interest or pleasure in almost all activities
Weight change Substantial unintentional weight gain or loss
Sleep disturbance Insomnia or hypersomnia nearly every day
Psychomotor problems Psychomotor agitation or retardation nearly every day
Lack of energy Fatigue or loss of energy nearly every day
Excessive guilt Feelings of worthlessness or excessive guilt nearly every day
Poor concentration Diminished ability to think or concentrate nearly every day
Suicidal ideation Recurrent thoughts of death or suicide
SOURCE: APA 2000

dosing and continuation of therapy for at least 6 to 12
months. For patients who relapse following drug dis-
continuation, or who have had prior episodes of de-
pression or anxiety, extended periods of treatment are
recommended. For those with multiple episodes, main-
tenance therapy is the standard of care (Fava 1994).
SSRI/SNRI OVERVIEW
Mechanism of action
The precise mechanism of action of SSRIs and SNRIs
remains undetermined but may be associated with al-
terations in gene expression that produce sustained
changes in the function of selected brain cells. Simply in-
creasing or reducing brain levels of serotonin is proba-
bly insufficient to produce an antidepressant response.
Newer agents have been developed, following the recog-
nition that balancing levels of serotonergic, noradrener-
gic, and other neurotransmitter activity may be required.
Efficacy, effectiveness, and cost considerations
Prospective, randomized, controlled trials help to es-
tablish the efficacy of antidepressants in reducing de-
pression symptom severity based on a standardized rat-
ing scale. For treatment of anxiety and mood disorders,
SSRIs and SNRIs appear no more efficacious than older
agents, such as tricyclic antidepressants (TCAs) and
monoamine oxidase (MAO) inhibitors (Williams 2000).
SSRIs and SNRIs are used much more widely, however,
in part because they are associated with lower rates of ad-
verse events (MacGillivray 2003) and better tolerated.
Although efficacy of these drugs has been demon-
strated in randomized controlled trials, these studies are
limited in their number of subjects and in their follow-
up duration, which is generally shorter than the length
of time antidepressants would be used in routine clini-
cal settings. Also, participants in psychiatric studies may
differ from typical patients in primary care settings.
Thus, it may be difficult to apply such findings to primary
care settings, where most antidepressants are prescribed.
Effectiveness is measured retrospectively based on
processes of care (e.g., guideline adherence) and clinical
outcomes achieved under naturalistic conditions and
across populations. SSRIs are most effective when used for
sufficient periods and at sufficient doses to achieve and
sustain full symptom remission. If used appropriately,
SSRIs generally are regarded as more effective than TCAs
and MAO inhibitors. SSRIs appear to be more effective
than TCAs for treating depression in women (Yonkers
2003) and have demonstrated effectiveness as first-line
therapy for postpartum depression (Wisner 2002).
Cost considerations are important in formulary de-
velopment but can be counterproductive if given more

26 P&T DIGEST

importance than effectiveness of a drug. For example,
health care costs associated with 6 months of treatment
with fluoxetine or a much less costly TCA (imipramine
or desipramine) were equivalent as demonstrated in a
randomized controlled trial. The higher cost of fluoxe-
tine was offset by reduced spending for outpatient visits
and inpatient services (Simon 1996). Therefore, even
though older agents are relatively inexpensive, SSRIs may
be more cost-effective (Weilburg 2003, Singletary 1997).
AGENTS FOR DEPRESSIVE DISORDERS
Initiation of treatment
No single SSRI or SNRI has demonstrated clear supe-
riority. Despite the apparent similarity between these
antidepressants, patients often will respond to one agent
but not another. As a result, there are no well-defined
guidelines for selecting the optimal agent for initial treat-
ment of depressive disorders. Various treatment algo-
rithms are based, at least partly, on cost. Likewise, many
drug benefit plans offer three tiers, with generics on the
first tier for no copayment or a minimal copayment.
Such a structure tends to steer patients and clinicians to-
ward TCAs and, more recently, toward generic SSRIs. In
clinical practice, physicians may choose to begin therapy
with agents marketed for the treatment of anxiety and
mood disorders for those patients with a combination of
symptoms. They also may start with agents that have
been useful for other family members with depression.
Although the practice is controversial, some physicians
choose agents based on the anticipated side-effect profile.
Thus, they will use agents less likely to produce anxiety or
insomnia for patients who complain of these symptoms,
and more “activating”agents, such as fluoxetine, for those
TABLE 1 Indications for SSRIs and SNRIs
Major Generalized Obsessive-
depressive anxiety compulsive Panic
Generic name disorder disorder disorder disorder disorder
Citalopram X
Escitalopram X X
Fluvoxamine
Fluoxetine (Prozac) X X
Fluoxetine (Sarafem)
Paroxetine X X
Paroxetine
(controlled release) X X
Sertraline X X
Venlafaxine X
Venlafaxine
(extended release) X X
SNRI=serotonin norepinephrine reuptake inhibitor; SSRI= selective serotonin reuptake inhibitor.
SOURCE: MANUFACTURERS’ PACKAGE INSERTS
P H A R M ACOT H E R A P Y
patients who are anergic. Some evidence suggests that
sustained-release formulations may produce less nausea
than their immediate-release formulations (DeVane
2003). Nausea, a relatively common medication side ef-
fect, may be associated with premature discontinuation,
which may support making sustained-release agents avail-
able for treatment initiation.
Treatment continuation
Many patients cannot tolerate side effects associated
with an initial antidepressant or may fail to respond to
it. Given that the goal of therapy is achieving remission,
such patients should be switched to other agents. This is
well illustrated by an open-label, naturalistic, 9-month
trial in which 573 depressed adults were randomly as-
signed to begin treatment with fluoxetine, paroxetine, or
sertraline (Kroenke 2001). The decision to initiate treat-
ment was based on a primary care physician’s judgment
that a patient had clinical depression warranting treat-
ment (as opposed to meeting criteria for a diagnosis). If
a patient did not respond adequately or tolerate the ini-
tial SSRI, the primary care physician was free to switch to
a different SSRI or an antidepressant from another class.
After 1 month, 13 percent of patients switched to an-
other antidepressant; after 3 months, 23 percent; 6
months, 32 percent; and 9 months, 40 percent. The ini-
tial SSRI the patient received at randomization made no
difference with respect to discontinuation or switching,
and each SSRI had a similar side-effect profile. The per-
centage of patients who remained on their initial SSRI for
the entire 9 months and scored 40 or higher on the 36-
item Short-Form Mental Component Summary was 37
percent in the fluoxetine group, 34 percent in the parox-
Post-
traumatic Premenstrual Social
stress dysphoric anxiety Bulimia
disorder disorder nervosa
X
X X
X
X X X X
X X
X X X X
DEPRESSION 27
 P H A R M ACOT H E R A P Y
etine group, and 37 percent in the sertraline group. At the
end of the study, 26 percent of the population met the cri-
teria for MDD, compared with 74 percent at baseline and
32 percent at 3 months.
Most patients require therapy of several weeks’ dura-
tion to exhibit a full response. At least 8 weeks of treat-
ment should be allowed to elapse before nonresponse to
fluoxetine is declared (Quitkin 2003). The American
College of Physicians–American Society of Internal Med-
icine reports that patients who have responded poorly or
not at all after 6 weeks of treatment with an adequate dose
of an antidepressant are unlikely to benefit from con-
tinued treatment at that dose.
Recurrences are common in some forms of unipolar
depression, necessitating a continuation or maintenance
phase of therapy. Although patients who feel better may
have difficulty adhering to continuation therapy, it is
important for them to do so because high rates of relapse
are observed upon drug discontinuation during main-
tenance therapy. Patients at relatively higher risk for re-
lapse and recurrence are those with a recurrent course of
depression, double depression (cycle between MDD and
dysthymic disorder), long duration of the index episode
(e.g., >2 years), and residual depressive symptoms (Fava
1994).
The optimal dose for maintenance therapy in depres-
sion is the same dose that led to remission of the de-
pressive symptoms (Frank 1993). In a double-blind study
of patients whose depression was stabilized with paroxe-
tine 40 mg, the recurrence rate was 51 percent among
those who were randomly assigned to paroxetine 20 mg
compared with 23 percent among those who continued
on paroxetine 40 mg (Franchini 1998). The guidelines of
the American College of Physicians–American Society of
Internal Medicine for the pharmacologic treatment of
acute major depression and dysthymia recommend con-
tinuing drug treatment at the same dose for at least 4
months after recovery or improvement (Snow 2000).
Relapse and recurrence occur despite antidepressant
treatment in 10 to 30 percent of patients (Fava 1994). To
minimize the risk of relapse, depressed patients who
achieve symptom remittance must continue therapy for
a substantial period of time — a minimum of 4 to 6
months. It has been shown, for example, that at least an
additional 26 weeks of fluoxetine treatment is necessary
to reduce the risk of relapse following an initial 12-week
course of fluoxetine (Reimherr 1998). Depending on the
number of recurrences, lifelong prophylactic therapy
may be warranted (Montgomery 1996).
Patients who experience a relapse or recurrence dur-
ing antidepressant treatment may respond to an increase
in dose. In a small study (N=18), depressed patients who
had recovered on fluoxetine 20 mg and then relapsed
28 P&T DIGEST
were treated with fluoxetine 40 mg, with 67 percent
showing a full response and 17 percent showing a par-
tial response (Fava 1995). During follow-up for a mean
of 4.7 months, 11 of the 18 patients maintained response.
In a larger trial (N=501), patients who responded to flu-
oxetine 20 mg were randomized to fluoxetine 20 mg
daily, fluoxetine 90 mg weekly, or placebo during a 25-
week continuation phase (Schmidt 2002). Patients who
relapsed during the continuation phase were offered a re-
sumption of fluoxetine 20 mg daily if they had been on
placebo, fluoxetine 40 mg daily if they had been on 20 mg
daily, and fluoxetine 90 mg twice weekly if they had been
on 90 mg once weekly. A total of 57 percent of the 40 mg
group and 72 percent of the 90 mg twice-weekly group
responded to the dose increase; 35 percent of patients ei-
ther failed to respond or relapsed after the dose increase.
Switching to another agent also can be an option to
treat relapsed or recurrent depression and may result in
reduced drug costs, fewer drug-drug interactions, im-
proved adherence, and greater convenience for the pa-
tient, as opposed to adding another agent (Marangell
2001). When a patient is switched to another SSRI, it is
preferable to base the dose of the new SSRI on that of the
original one, instead of subjecting the patient to a
washout period that may induce discontinuation effects.
Persistence of symptoms after the end of a depressive
episode points to a patient at increased risk for another
episode of depression progressing from the residual
symptoms (Nierenberg 2003). Long-term drug therapy
is recommended for patients at high risk of relapse or re-
currence. Augmenting pharmacotherapy with psycho-
therapy also is effective in reducing the risk of new
episodes of depression, as is psychotherapy alone.
Nonadherence to the drug regimen — anything less
than 80 to 85 percent of the prescribed medication —
should be suspected whenever relapse occurs during the
continuation phase of therapy for patients who had
achieved full remission (Thase 2003).
AGENTS FOR ANXIETY DISORDERS
General anxiety disorder
For patients with GAD, SSRIs and SNRIs are sup-
planting benzodiazepines as first-line treatment (Sramek
2002). GAD is commonly comorbid with depression,
and an SSRI often can be effective for both disorders,
sparing the patient the need to take a second drug.
Obsessive-compulsive disorder
In the treatment of OCD, a combination of drug and
cognitive-behavioral therapies (CBT) is the most effec-
tive strategy for restoring social functioning, even if com-
plete remission of symptoms is rarely achieved (Jenike
2004). Drug therapy should be initiated before CBT is
 P H A R M ACOT H E R A P Y

added. An SSRI appears to be a better choice than a TCA Panic disorder

for initial therapy, but efficacy within the SSRI class
seems to be similar. A trial of 10 to 12 weeks (longer than
might be employed for MDD) may be necessary, at a
higher dose than for MDD. If a patient fails two or three
consecutive SSRIs, a trial of clomipramine is warranted.
For treatment of panic disorder with or without agora-
phobia, SSRIs have been identified by an international
consensus group as the drugs of first choice, used for 12
to 24 months and discontinued over the course of 4 to 6
months (Ballenger 1998). A 10-year longitudinal study

TABLE 2 Selected pharmacologic properties of SSRIs and SNRIs

Linear Major metabolic Possible

Generic % protein- Tmax pharmaco- pathways; drug-drug
name bound (hours) Half-life kinetics excretion interactions*
Citalopram 80 3–4 35 hours Yes 3A4, 2C19; renal MAO inhibitors
clearance, 20%
Escitalopram 56 5 27–32 hours Yes 3A4, 2C19; MAO inhibitors
renal clearance, 7%
Fluoxetine 94 6–8 1–3 days after acute No 2D6; MAO inhibitors
administration; 4–6 primarily urine benzodiazepines
days after chronic NSAIDs/aspirin
administration thioridazine
phenytoin
active metabolite:
carbamazepine
4–16 days after
warfarin
acute and chronic
administration

Fluvoxamine 77 2-8 15 hours No Oxidative MAO inhibitors

demethylation tacrine
and deamination; metoprolol
urine, 94% propranolol
Paroxetine 95 5.2 20 hours No 2D6; urine, 64% MAO inhibitors
NSAIDs/aspirin
warfarin
tryptophan
thioridazine
Paroxetine 95 6–10 15–20 hours No 2D6; urine, 64% MAO inhibitors
(controlled NSAIDs/aspirin
release) warfarin
tryptophan
thioridazine
Sertraline 98 6–8 26 hours Yes 3A3/4; urine, MAO inhibitors
40–45% warfarin
Venlafaxine 27 2 5 hours Yes 2D6; urine, 87% MAO inhibitors
active metabolite:
11 hours
Venlafaxine 27 5.5 5 hours Yes 2D6; urine, 87% MAO inhibitors
(extended
release)

*This listing is not inclusive. Consult product labeling for details.

MAO inhibitor=monoamine oxidase inhibitor; NSAID=nonsteroidal antiinflammatory drug;Tmax=time to maximum (peak) concentration.
SOURCE: MANUFACTURERS’ PACKAGE INSERTS

DEPRESSION 29
 P H A R M ACOT H E R A P Y
suggests that SSRIs may be no more effective for panic
disorder than common benzodiazepines (Bruce 2003).
Still, the SSRIs deserve consideration on the basis of
greater safety and tolerability, minimal potential for
abuse, and comorbidity with other disorders for which
they have been shown to be efficacious (Zamorski 2002).
Patients with panic disorder tend to need SSRI dosages
that are higher than for other conditions. The starting
dose, however, is usually half the dose for depression.
Posttraumatic stress disorder
On the basis of their safety and tolerability, SSRIs also are
first-line drugs for management of PTSD (Yehuda 2002).
If there is no response after an 8-week trial of an SSRI, a
consensus group recommends that treatment be switched
to venlafaxine or nefazodone. The nefazodone label cau-
tions about liver toxicity, and recommends discontinu-
ance if serum AST or serum ALT reaches 3 times normal.
Social anxiety disorder
Paroxetine, sertraline, venlafaxine, and high-potency
benzodiazepines are medications of choice among those
indicated. Developing an effective treatment plan de-
pends on making a diagnostic distinction between gen-
eralized SAD and nongeneralized SAD, as therapeutic op-
tions tend to differ (see article by Raj, page 52).
CHARACTERISTICS OF SPECIFIC SSRIs/SNRIs
Information about metabolism may be helpful for de-
termining which drug might be more appropriate for pa-
tients taking concomitant drugs that may interact with the
same metabolic pathways. Table 2 (page 29) lists selected
pharmacologic properties of SSRIs and SNRIs; table 3
(page 31) lists their dosing schedules. Concurrent use of
SSRIs/SNRIs and MAO inhibitors is contraindicated, but
the recommended washout interval between agents varies.
Fluoxetine
Fluoxetine (Prozac, Sarafem) was introduced in the
United States in 1987. Except for having different trade
names and capsule colors, Prozac and Sarafem are iden-
tical. Fluoxetine is available as a tablet, capsule (Pulvule),
or oral solution for daily use, or as a delayed-release 90
mg capsule for weekly use.
Both fluoxetine and its active metabolite norfluoxetine
have extremely long elimination half-lives, persisting in
the body for many weeks after discontinuation. Fluoxe-
tine therefore tapers naturally. There should be at least 5
weeks between discontinuation of fluoxetine and initia-
tion of an MAO inhibitor or thioridazine, however.
In clinical trials, the most commonly reported side ef-
fects of fluoxetine were nausea (22 percent of subjects
given fluoxetine vs. 9 percent of subjects given placebo)
30 P&T DIGEST
and insomnia (19 vs. 10 percent, with fluoxetine and
placebo, respectively).2
Sertraline
Sertraline (Zoloft) was the second SSRI to enter the
U.S. market, earning FDA approval in 1991. It is available
as coated tablets or oral concentrate. The concentrate is
12 percent alcohol, so it should not be given to patients
using disulfiram (Antabuse).
Sertraline should not be used in combination with an
MAO inhibitor, and at least 14 days should elapse be-
tween MAO inhibitor discontinuation and sertraline ini-
tiation, or vice versa.
As with fluoxetine, nausea was the most frequently re-
ported adverse effect of sertraline (25 vs. 11 percent,
with sertraline and placebo, respectively). Insomnia was
the next most commonly reported side effect (21 vs. 11
percent, with sertraline and placebo, respectively).
Paroxetine
Paroxetine (Paxil) first became available in 1992, and
a controlled-release version (Paxil CR) was approved in
1999. The tablets of controlled-release paroxetine contain
a polymeric matrix that degrades at a controlled rate
over the course of 4 to 5 hours. These tablets also have
an enteric coating that delays drug release until the tablet
has left the stomach.
Because of the 2D6 metabolism, paroxetine should not
be used in combination with thioridazine, which also is
metabolized by 2D6. Concurrent use of paroxetine could
result in elevated plasma levels of thioridazine, which is
associated with an apparently dose-related increase in the
QTc interval.
Paroxetine should not be used in combination with an
MAO inhibitor. Paroxetine treatment should not be ini-
tiated within 14 days of discontinuing an MAO inhibitor,
and at least 2 weeks should elapse between discontinu-
ation of paroxetine and initiation of an MAO inhibitor.
When paroxetine treatment is discontinued, the dose
should be reduced gradually to reduce the risk of dis-
continuation symptoms.
The most common side effect of immediate-release
paroxetine was nausea (26 vs. 9 percent, with paroxetine
and placebo, respectively). Ejaculation disorder (26 vs. 1
percent, with paroxetine and placebo, respectively) was
most commonly reported by subjects taking controlled-
release paroxetine.
Venlafaxine
Venlafaxine (Effexor) was the first SNRI to become
2 The treatment-emergent side effects reported herein are based
on placebo-controlled trials cited on manufacturers’ package in-
serts. Comparisons shown are versus placebo.
 P H A R M ACOT H E R A P Y

available in the United States, receiving FDA approval in
1993. Extended-release venlafaxine (Effexor XR) was
approved in 1997. Release of drug by the extended-
release product is controlled by diffusion through the
coating membrane and is not pH-dependent.
At least 14 days should pass between discontinuation
of an MAO inhibitor and initiation of venlafaxine, and
an interval of at least 7 days should separate venlafaxine
discontinuation and MAO inhibitor initiation.
In clinical trials, a mean increase in pulse rate of 3 beats
per minute was observed among patients receiving ven-
lafaxine, versus no change among patients receiving
placebo. In addition, a dose-related increase in diastolic
blood pressure was observed. For all doses, the overall
mean increases ranged from 0.7 to 2.5 mm Hg, versus de-
creases of 0.9 to 3.8 mm Hg for placebo. At dosages less

TABLE 3 Dosing and administration of SSRIs and SNRIs

Generic Products Initial Therapeutic

name available dose dose Administration
Citalopram Tablets (10, 20, 40 mg) MDD: 20 mg MDD: 40 mg Once daily, morning or evening,
Oral solution (10 mg/5 mL) with or without food
Escitalopram Tablets (5, 10, 20 mg) MDD: 10 mg MDD: 10 mg Once daily, morning or evening,
Oral solution (5 mg/5 mL) GAD: 10 mg GAD: 10 mg with or without food
Fluoxetine Tablets (10 mg) MDD: 20 mg MDD: 20–80 mg Initially, once daily in morning,
Capsules (10, 20, 40 mg) Bulimia: 60 mg Bulimia: 60 mg with or without food
Oral solution (20 mg/5 mL) OCD: 20 mg OCD: 20–60 mg >20 mg, once daily in morning
Delayed-release capsules Panic: 10 mg Panic: 20–60 mg or twice daily (morning and
(90 mg) PMDD: 20 mg PMDD: 20 mg noon)
Fluvoxamine Tablets (25, 50,100 mg) OCD: 50 mg OCD: 50–200 mg Initially, single dose, at bedtime;
>100 mg, divided dose, with
larger portion given at bedtime
Paroxetine Scored tablets (10, 20 mg) MDD: 20 mg MDD: 20–50 mg Single dose, preferably the
Tablets (30, 40 mg) GAD: 20 mg GAD: 20 mg morning, with or without food
Oral suspension OCD: 20 mg OCD: 20–60 mg
(10 mg/5 mL) Panic: 10 mg Panic: 10–60 mg
PTSD: 20 mg PTSD: 20–40 mg
SAD: 20 mg SAD: 20 mg
Paroxetine Coated tablets (12.5, 25.0, MDD: 25.0 mg MDD: 25.0–62.5 mg Single dose, preferably the
(controlled 37.5 mg) Panic: 12.5 mg Panic: 12.5–75.0 mg morning, with or without food
release) PMDD: PMDD:12.5–25.0 mg
12.5 mg SAD: 12.5–37.5 mg
SAD: 12.5 mg
Sertraline Scored tablets (25, 50, MDD: 50 mg 50–200 mg Once daily, morning or evening,
100 mg) OCD: 50 mg with or without food
Coated tablets (25, 50, Panic: 25 mg
100 mg) PMDD: 50 mg
Oral concentration PTSD: 25 mg
(20 mg/mL) SAD: 25 mg
Venlafaxine Tablets (25.0, 37.5, 50.0, 75.0, MDD: 75 mg 75–375 mg Two or three divided doses,
100.0 mg) with food
Venlafaxine Capsules (37.5, 75.0 mg) MDD: 75 mg MDD: 75–375 mg
(extended GAD:37.5–75.0 mg GAD: 75–225 mg Single dose, with food, morning
release) SAD:37.5–75.0 mg SAD: 75–225 mg or evening

GAD=generalized anxiety disorder; MDD=major depressive disorder; OCD=obsessive-compulsive disorder; PMDD=premenstrual dysphoric disorder;
PTSD=posttraumatic stress disorder; SAD=social anxiety disorder.
SOURCE: MANUFACTURERS’ PACKAGE INSERTS

DEPRESSION 31
 P H A R M ACOT H E R A P Y
than 100 mg a day, the incidence of a sustained elevation
in supine diastolic blood pressure was 3 percent (placebo,
2 percent); 101 to 200 mg a day, 5 percent; 201 to 300 mg
a day, 7 percent; more than 300 mg a day, 13 percent. Reg-
ular monitoring of blood pressure thus is advised for pa-
tients receiving venlafaxine.
When venlafaxine is discontinued, if patients have re-
ceived more than 1 week of treatment, the dose should
be tapered to reduce the risk of discontinuation symp-
toms. If the treatment period exceeds 6 weeks, the dose
should be tapered gradually over the course of 2 weeks.
Nausea was the most prevalent adverse effect of both
the immediate- and extended-release forms of venla-
faxine (37 vs. 11 percent and 31 vs. 12 percent, with ven-
lafaxine and placebo, respectively).
Fluvoxamine
Fluvoxamine (Luvox) was the first SSRI to receive an
indication for OCD. Although it lacks an FDA-approved
indication for MDD, fluvoxamine often is used to treat
depression, much as other SSRIs commonly are used
without regard to their official indications.
Unlike other SSRIs, which may be given as a single
dose, it is advised that fluvoxamine doses exceeding 100
mg be given in two divided doses, with the larger dose (if
applicable) being given at bedtime.
Two weeks should be allowed to elapse between the
discontinuation of an MAO inhibitor and the initiation
of fluvoxamine, and vice versa.
The most commonly reported adverse effects with
fluvoxamine were nausea (40 vs. 14 percent, with flu-
voxamine and placebo, respectively) and somnolence
(22 vs. 8 percent with fluvoxamine and placebo, respec-
tively).
Citalopram
Citalopram (Celexa) is a racemic mixture, containing
the R- and S-enantiomers in equal proportion. The S-
enantiomer is responsible for most of citalopram’s inhi-
bition of serotonin reuptake.3
MAO inhibitors should not be administered with citalo-
pram or escitalopram, and at least 14 days should elapse
between discontinuation of an MAO inhibitor and initi-
ation of citalopram or escitalopram therapy, and vice versa.
The most common side effects of citalopram are nau-
sea (21 vs. 14 percent, with citalopram and placebo, re-
spectively) and dry mouth (20 vs. 14 percent, with citalo-
pram and placebo, respectively).
3 Enantiomers, or stereoisomers, are mirror images of molecules
with the same molecular formula. R (from the Latin rectus) in-
dicates the right-handed version, S (from the Latin sinister), the
left. Because they have different spatial configurations, R- and S-
enantiomers may have very different pharmacologic properties.
32 P&T DIGEST
Escitalopram
Escitalopram (Lexapro) was the last SSRI to receive
FDA approval, in 2002. As its generic name suggests, es-
citalopram consists entirely of the therapeutically active
S-enantiomer of citalopram. One consequence of isolat-
ing the active isomer is that escitalopram is more potent
than citalopram, with 10 mg of escitalopram being the
equivalent of about 40 mg of citalopram. Escitalopram 10
mg has been shown to be as efficacious as citalopram 40
mg in the treatment of MDD (Burke 2002). In another
study, a greater percentage of moderately to severely de-
pressed patients responded to treatment with escitalo-
pram than did those receiving citalopram (P=.021) (Lep-
ola 2003). Escitalopram, along with citalopram, has been
shown to be efficacious in the treatment of panic disor-
der (Stahl 2003), but neither agent has yet received such
an indication.
Escitalopram offers a side-effect profile that is similar
to citalopram (Lepola 2003, Burke 2002).
In fixed-dose trials, the incidence rate of some ad-
verse events increased in a dose-related fashion. For ex-
ample, insomnia was reported in 7 percent of patients re-
ceiving escitalopram 10 mg but 14 percent of those
receiving escitalopram 20 mg; diarrhea, 6 and 14 percent;
dry mouth, 4 and 9 percent; somnolence, 4 and 9 percent;
dizziness, 4 and 7 percent, respectively.
SUMMARY
The SSRIs and SNRIs are used commonly for treating
mood and anxiety disorders. They are effective when
used properly; proper use entails providing a patient
with an adequate dose for an adequate period of time be-
fore deciding that an agent is ineffective. No evidence fa-
vors one SSRI or SNRI over another as initial therapy for
a given anxiety or mood disorder. Physicians and patients
should be free to begin treatment with any SSRI or SNRI
deemed appropriate and to switch to another SSRI or
SNRI whenever necessary.
REFERENCES
APA (American Psychiatric Association). Diagnostic and Statisti-
cal Manual of Mental Disorders, Fourth Edition, Text Revi-
sion (DSM-IV-TR). Arlington, Va.: American Psychiatric
Publishing. 2000.
Ballenger JC, Davidson JR, Lecrubier Y, et al. Consensus state-
ment on panic disorder from the International Consensus
Group on Depression and Anxiety. J Clin Psychiatry.
1998;59(suppl 8):47–54.
Bruce SE, Vasile RG, Goisman RM, et al. Are benzodiazepines
still the medication of choice for patients with panic dis-
order with or without agoraphobia? Am J Psychiatry.
2003;160:1432–1438.
Burke WJ, Gergel I, Bose A. Fixed-dose trial of the single isomer
SSRI escitalopram in depressed outpatients. J Clin Psych-
iatry. 2002;63:331–336.
Celexa (citalopram hydrobromide) package insert. St. Louis:
Forest Pharmaceuticals. 2002.

DeVane CL. Immediate-release versus controlled-release formu-
lations: pharmacokinetics of newer antidepressants in rela-
tion to nausea. J Clin Psychiatry. 2003;64(suppl 18):14–19.
Druss BG, Rosenheck RA, Sledge WH. Health and disability
costs of depressive illness in a major U.S. corporation. Am
J Psychiatry. 2000;157:1274–1278.
Effexor (venlafaxine HCl) package insert. Philadelphia: Wyeth
Pharmaceuticals. 2003.
Effexor XR (venlafaxine HCl) extended-release capsules package
insert. Philadelphia: Wyeth Pharmaceuticals. 2003.
Fava M, Kaji J. Continuation and maintenance treatments for
major depressive disorder. Psychiatr Ann. 1994;24:281–290.
Fava M, Rappe SM, Pava JA, et al. Relapse in patients on long-
term fluoxetine treatment: response to increased fluoxetine
dose. J Clin Psychiatry. 1995;56:52–55.
Franchini L, Gasperini M, Perez J, et al. Dose-response efficacy
of paroxetine in preventing depressive recurrences: a ran-
domized, double-blind study. J Clin Psychiatry. 1998;
59:229–232.
Frank E, Kupfer DJ, Perel JM, et al. Comparison of full-dose ver-
sus half-dose pharmacotherapy in the maintenance treat-
ment of recurrent depression. J Affect Disord. 1993;27:
139–145.
Jenike MA. Obsessive-compulsive disorder. N Engl J Med.
2004;350:259–265.
Kroenke K, West SL, Swindle R, et al. Similar effectiveness of
paroxetine, fluoxetine, and sertraline in primary care: a
randomized trial. JAMA. 2001;286:2947–2955.
Lepola UM, Loft H, Reines EH. Escitalopram (10–20 mg/day) is
effective and well tolerated in a placebo-controlled study in
depression in primary care. Int Clin Psychopharmacol.
2003;18:211–217.
Lexapro (escitalopram oxalate) package insert. St. Louis: Forest
Pharmaceuticals. 2003.
Luvox (fluvoxamine maleate) package insert. Brussels, Belgium:
Solvay Pharmaceuticals. 2001.
MacGillivray S, Arroll B, Hatcher S, et al. Efficacy and tolerabil-
ity of selective serotonin reuptake inhibitors compared
with tricyclic antidepressants in depression treated in pri-
mary care: systematic review and meta-analysis. BMJ.
2003;326:1014.
Marangell LB. Switching antidepressants for treatment-resistant
major depression. J Clin Psychiatry. 2001;62(suppl
18):12–17.
Montgomery SA. Efficacy in long-term treatment of depression.
J Clin Psychiatry. 1996;57(suppl 2):24–30.
Nierenberg AA, Petersen TJ, Alpert JE. Prevention of relapse and
recurrence in depression: the role of long-term pharmaco-
therapy and psychotherapy. J Clin Psychiatry.
2003;64(suppl 15):13–17.
Paxil (paroxetine hydrochloride) package insert. Research Trian-
gle Park, N.C.: GlaxoSmithKline. 2003.
P H A R M ACOT H E R A P Y
Paxil CR (paroxetine hydrochloride) controlled-release tablets
package insert. Research Triangle Park, N.C.: GlaxoSmith-
Kline. 2003.
Prozac (fluoxetine hydrochloride) package insert. Indianapolis:
Eli Lilly and Co. 2003.
Quitkin FM, Petkova E, McGrath PJ, et al. When should a trial of
fluoxetine for major depression be declared failed? Am J
Psychiatry. 2003;160:734–740.
Reimherr FW, Amsterdam JD, Quitkin FM, et al. Optimal length
of continuation therapy in depression: a prospective assess-
ment during long-term fluoxetine treatment. Am J Psychia-
try. 1998;155:1247–1253.
Schmidt ME, Fava M, Zhang S, et al. Treatment approaches to
major depressive disorder relapse. Part 1: dose increase.
Psychother Psychosom. 2002;71:190–194.
Serzone (nefazodone HCl) tablets package insert. Princeton,
N.J.: Bristol-Myers Squibb. 2003.
Simon GE, VonKorff M, Heiligenstein JH, et al. Initial anti-
depressant choice in primary care. Effectiveness and cost
of fluoxetine vs tricyclic antidepressants. JAMA. 1996;
275:1897–1902.
Singletary T, North DS, Weiss M, Marman G. A cost-effective ap-
proach to the use of selective serotonin reuptake inhibitors
in a Veterans Affairs Medical Center. Am J Manag Care.
1997;3:125–129.
Snow V, Lascher S, Mottur-Pilson C. Pharmacologic treatment of
acute major depression and dysthymia. American College
of Physicians–American Society of Internal Medicine. Ann
Intern Med. 2000;132:738–742.
Sramek JJ, Zarotsky V, Cutler NR. Generalised anxiety disorder.
Treatment options. Drugs. 2002;62:1635–1648.
Stahl SM, Gergel I, Li D. Escitalopram in the treatment of panic
disorder: a randomized, double-blind, placebo-controlled
trial. J Clin Psychiatry. 2003;64:1322–1327.
Thase ME. Achieving remission and managing relapse in depres-
sion. J Clin Psychiatry. 2003;64(suppl 18):3–7.
Weilburg JB, O’Leary KM, Meigs JB, et al. Evaluation of the ade-
quacy of outpatient antidepressant treatment. Psychiatr
Serv. 2003;54:1233–1239.
Williams JW Jr, Mulrow CD, Chiquette E, et al. A systematic review
of newer pharmacotherapies for depression in adults: evi-
dence report summary. Ann Intern Med. 2000;132:743–756.
Wisner KL, Parry BL, Piontek CM. Postpartum depression. N
Engl J Med. 2002;347:194–199.
Yehuda R. Current concepts. Post-traumatic stress disorder. N
Engl J Med. 2002;346:108–114.
Yonkers KA. Special issues related to the treatment of depression
in women. J Clin Psychiatry. 2003;64(suppl 18):8–13.
Zamorski MA, Albucher RC. What to do when SSRIs fail: eight
strategies for optimizing treatment of panic disorder. Am
Fam Physician. 2002;66:1477–1484.
Zoloft (sertraline HCl) package insert. New York: Pfizer. 2003.
DEPRESSION 33
 CO S T - E F F E C T I V E N E S S CO S T - E F F E C T I V E N E S S
Pharmacoeconomic Evaluation
Of Antidepressant Therapies
MATTHEW W. SARNES, PHARMD,1 AND LAURA E. FRANKUM, PHARMD2
Applied Health Outcomes
SUMMARY
With today’s emphasis in the managed care
environment on evidentiary literature, new
antidepressants must provide economic evi-
dence of their value compared with previ-
ously available agents.Various pharmaco-
economic models have shown newer
antidepressant agents to be cost-effective
relative to older agents.The authors review
methodologies and published results.
S
ince its inception in the early 1980s, pharmaco-
economics as an industry and a science has
evolved and flourished. Pharmacoeconomics
traditionally focused solely on evaluating the
economic impact of pharmaceutical products to the
health care system; however, it has matured into a disci-
pline that compares the clinical and economic outcomes
associated with various pharmacotherapies against the
inherent costs of providing those therapies. The goal of
these types of assessments is to determine if the value de-
1 Matthew W. Sarnes, PharmD, is a senior consultant with
Applied Health Outcomes, a health outcomes research and
consulting firm in Palm Harbor, Fla. His responsibilities in-
clude the marketing, development, and implementation of
services that measure the value of pharmaceuticals and other
health care technologies. The author of several scientific pub-
lications, Sarnes has been responsible for the design and
management of economic models, prospective postmarket-
ing studies, database analyses, productivity assessments,
quality-improvement programs, and formulary dossiers. He
received his PharmD from the University of Pittsburgh.
2 Laura E. Frankum, PharmD, is an outcomes research fel-
low at Applied Health Outcomes. Her responsibilities include
formulary dossier development, pharmacoeconomic mod-
eling, manuscript development, and content development
and dissemination of data for disease management pro-
grams. She received her PharmD from Creighton University.
34 P&T DIGEST
rived from a pharmaceuti-
cal agent justifies its acqui-
sition and administration
costs and the costs that are
associated with potential
side effects. In addition,
pharmacoeconomic analy-
ses provide health care de-
cision makers with data to
help them determine the
value of a pharmaceutical
agent versus another.
Drug-acquisition costs
and the vast number of
drug products that are as- Matthew W. Sarnes, PharmD
sociated with depression
have made antidepressant pharmacoeconomic data para-
mount to formulary committees. Pharmacoeconomic
literature regarding depression management during the
past 2 decades has focused primarily on three questions:
Are neurotransmitter receptor-specific antidepressants
cost-effective alternatives to older agents? Are there dif-
ferences in clinical and economic outcomes between
newer antidepressants? Does compliance with guide-
lines regarding recommended length of antidepressant
therapy translate into beneficial economic outcomes?
Five distinct methodologies (randomized clinical tri-
als, prospective naturalistic inquiries, meta-analyses, de-
cision analytic models, and retrospective database analy-
ses) have been employed to answer these questions. Each
of these study designs has a role in the assessment
process. In fact, several studies using various method-
ologies typically are needed at different points during a
pharmaceutical product’s life cycle to assess its value.
Herein, each methodology will be described, inherent
strengths and weakness will be considered, and relevant
examples from the literature will be presented.
RANDOMIZED CLINICAL TRIALS
Randomized clinical trials (RCTs) with an economic
component often are considered the gold standard of
pharmacoeconomic analyses by clinicians. Random as-
 CO S T - E F F E C T I V E N E S S

signment of interventions to sam-

ples of demographically similar TABLE 1 Costs and efficacy for TCAs and SSRIs
populations inherently controls
Amitriptyline Citalopram Fluoxetine P value
for known and unknown con-
founding variables. In addition, Sample size 31 29 30
patient and investigator blinding Median total direct costs* $1,067 $1,128 $1,083 0.97
in these studies minimizes poten- Median number of
tial biases. Conclusions from RCTs hospitalization-free days 140 155 152 0.18
may provide the strongest evi- TCAs=tricyclic antidepressants; SSRIs=selective serotonin reuptake inhibitors.
*Costs converted from Czech Crowns to 1997 U.S. dollars.
dence for causality, as temporality
SOURCE: HOSAK 2000
is certain and internal validity is
maximized. Nevertheless, internal
validity often is achieved at the expense of external va- at 4 and 6 months, using the Montgomery-Asberg De-
lidity (ability to generalize to a broader population), an pression Rating Scale (MADRS) and Clinical Global Im-
inherent limitation of RCTs (Frank 2001). In addition, pressions (CGI) severity scores for clinical outcomes,
RCTs are, by design, time and resource-consuming, thus and the Functional Status Questionnaire (FSQ) for qual-
restricting their usage. ity-of-life measures. Direct costs, from a third-party
Data from RCTs assessing the economic impact of payer perspective, were determined from medical-service
antidepressant comparators generally are limited, and re- use. Indirect costs, from the societal perspective, were de-
view articles have reported that the available data lack termined from work absence and productivity loss.
conclusive evidence (Skaer 2000). Hosak (2000) con- Significant clinical and quality-of-life improvements
ducted a prospective, open, intent-to-treat trial assessing were observed in both groups compared to baseline, but
the cost and efficacy of antidepressants in the Czech Re- there were no between-group differences. Fluoxetine pa-
public. This analysis was a continuation of an RCT as- tients utilized more medical resources and incurred more
sessing the efficacy and tolerability of amitriptyline, costs from the societal and third-party payer perspectives
citalopram, and fluoxetine in hospitalized patients. In (Table 2); the statistical significance of cost difference was
this analysis, clinical and economic endpoints for pa- not reported, however. The validity of this trial was ques-
tients following hospital discharge were collected via tioned subsequently, as antidepressant therapy costs were
questionnaires. The questionnaires were sent to outpa- not considered and two patients from the sertraline group
tient psychiatrists, who reported utilization of psycho- who attempted suicide were excluded (Frank 2001).
tropic medications, outpatient visits, and rehospitaliza-
tions. At the conclusion of the study, the investigators PROSPECTIVE NATURALISTIC INQUIRIES
found no significant differences in efficacy, defined as the As mentioned previously, RCTs apply strict inclusion
number of hospitalization-free days, between the treat- and exclusion criteria to their sample population to best
ment groups. In addition, while drug-acquisition costs determine a causal relationship between a pharmaceuti-
were higher for the selective serotonin reuptake in- cal intervention and the outcomes of interest. Demon-
hibitors (SSRIs), total medical costs were not signifi- stration of such a causal relationship establishes the effi-
cantly different between the three treatment arms (Table cacy of the studied pharmaceutical interventions. While
1). The authors concluded that there is no advantage in RCT designs are a necessary first step in determining the
restricting the use of SSRIs, based on their comparable value of an intervention, patients in RCTs often have lit-
cost and superior safety and tolerability profile. tle comorbidity, are not taking medications that could in-
In addition to assessing pharma-
coeconomic differences between TABLE 2 Efficacy and costs for fluoxetine compared with sertraline
newer and older antidepressant
therapies, RCT designs have been Fluoxetine Sertraline P value
used to evaluate clinical and eco- Sample size 115 116
nomic differences between SSRIs. Mean decrease in MADRS score 17.9 17.9 NS
Boyer (1998) conducted a double- Third party payer costs* $643 $585 NR
blind, randomized trial of 231 pa- Societal costs* $1,735 $1,551 NR
tients initiating antidepressant MADRS=Montgomery-Asberg Depression Rating Scale; NS=not significant; NR=not reported.
therapy with sertraline and fluox- *Originally reported in French francs but converted here to nearest whole U.S. dollar. At the time of the
study, 1.00 FF=U.S. $0.1993.
etine within primary care clinics
in France. Outcomes were assessed SOURCE: BOYER 1998

DEPRESSION 35
 CO S T - E F F E C T I V E N E S S
terfere with the outcome of interest, and generally are
more compliant with the treatment regimen as a result of
routine follow-up and observation. This homogeneity
often prevents application of these results to the general
population. Prospective naturalistic trials address this
issue through diminished enrollment criteria and less in-
teraction with study subjects. This not only provides the
study a more diverse population, but also allows obser-
vation of a population that is inclined to follow natural
behavior patterns due to reduced clinician contact with
the population. External validity, however, is gained at the
expense of internal validity, as confounders are difficult
to control for in a heterogeneous population.
To date, only one naturalistic prospective trial assess-
ing costs has been conducted among antidepressants.
This analysis initially was conducted through 1 year and
was subsequently extended to obtain data at 2 years
(Simon 1996, Simon 1999). Four hundred seventy-one
adults in primary care clinics of a staff-model HMO
were randomized to initiate antidepressant therapy with
desipramine, fluoxetine, or imipramine. Patients were as-
sessed for clinical and cost outcomes at 6, 9, 12, 18, and
24 months. Clinical outcomes were measured using the
Hamilton Depression Rating Scale (HAM-D) and de-
pression subscale of the Hopkins Symptom Checklist,
and quality of life was assessed by the Medical Outcomes
Study Short Form 36 (SF-36) Health Survey. The HMO’s
accounting data were used to determine cost outcomes.
At the 1- and 2-year endpoints, fluoxetine patients were
significantly more likely to continue with their initial
antidepressant agent, and adverse effects were the most
commonly reported reason for change or discontinua-
tion of index agent. No significant differences in quality-
of-life outcomes or efficacy outcomes were detected be-
tween the three treatment arms. Although antidepressant
drug costs were higher for patients taking fluoxetine,
total medical costs essentially were identical between the
three treatment arms. The authors concluded that initial
selection of fluoxetine or a tricyclic antidepressant (TCA)
leads to similar outcomes, and patients initiated on flu-
oxetine are more likely to continue therapy.
META-ANALYSES
RCT and prospective naturalistic trials typically are de-
signed to answer a specific research question or set of re-
search questions. Often, the results or trends observed in
these trials cause clinicians and researchers to generate
additional related hypotheses that the original trial was
not designed to investigate. Meta-analyses provide a way
to investigate some of these hypotheses by quantitatively
and systematically aggregating results from several sim-
ilarly designed trials. This technique is particularly use-
ful when there are many small studies regarding a par-
36 P&T DIGEST
ticular topic or when the results of the available studies
are heterogeneous.
The most rigorous method of conducting a meta-
analysis is to aggregate the raw data from the individual
trials of interest, but these data sets typically are difficult
to obtain. Therefore, investigators most often use data
available from published literature. These data may be
subject to publication bias, as results are published se-
lectively; therefore, conclusions drawn from these types
of meta-analysis must be thoroughly evaluated. In addi-
tion, the results of meta-analyses are limited by the level
of external validity provided by the studies employed in
the meta-analysis.
To our knowledge, meta-analyses assessing economic
endpoints in depression have not been conducted in the
area of depression. It is likely that this is due to the lim-
ited amount of pharmacoeconomic RCT and prospec-
tive naturalistic inquiry data available to generate such
an analysis. Nonetheless, meta-analyses have been con-
ducted regarding clinical outcomes, particularly with re-
spect to discontinuation rates due to adverse events. Four
meta-analyses have concluded that significantly fewer
patients discontinue SSRIs resulting from adverse events
as compared with placebo or TCA (Montgomery 1994,
1995; Anderson 1995). These results have been used to
fuel the debate about selection of a cost-effective anti-
depressant pharmacotherapy. Arguably, lower discon-
tinuation rates due to adverse events observed with SSRIs
should translate to a cost benefit; the validity of the re-
sults ascertained in the aforementioned analyses has
been questioned, however (Gallo 1999, Hotopf 1996,
Freemantle 2000).
DECISION ANALYTIC MODELS
Because economic data often are not collected in clini-
cal trials, decision analytic models provide a technique
to estimate the cost impact of the clinical outcomes as-
sociated with new antidepressant therapies. In addition,
modeling may be used to estimate and compare the cost-
effectiveness of two or more antidepressants that have not
been studied in a head-to-head clinical trial or evaluated
in a naturalistic setting. These models are constructed by
populating decision trees or mathematically based com-
puter simulators (e.g., Markov models) with event data
from clinical trials and real-world cost data. An advan-
tage of this design is that it allows researchers to simu-
late different treatment patterns (e.g., switching thera-
pies, dose titration), identify cost and outcome drivers,
and extrapolate data to time frames that are not practi-
cal in clinical trials. When evaluating a decision analytic
model, it is important for the user to determine whether
the construct, assumptions, and time frame are relevant
to their environment.

TABLE 3 Cost, efficacy, and quality-of-life outcomes at 9 and 21 months
Direct costs Lost productivity
9 months
Termination group* $474 $909
Prolongation group† $1,276 $304
21 months
Termination group $4,681 $4,282
Prolongation group $3,831 $1,512
QALYs=quality-adjusted life years. *Termination group: patients who went into remission and discontinued pharmacologic therapy. †Prolongation
group: patients who went into remission and continued with their index selective serotonin reuptake inhibitor.
SOURCE: NUIJTEN 2001
The construct of a model typically depends on which
perspective is taken when developing the model. The so-
cietal perspective is the most inclusive one and attempts
to assess all costs (both direct and indirect) associated
with an intervention or disease state. Models that include
only direct health care costs or only indirect health care
costs, however, such as productivity losses, may be more
useful in assessing outcomes from the standpoint of the
insurance payer or employer.
Once it is determined that the construct is relevant to
the user’s health care environment, it is essential to eval-
uate the validity of the inputs and critical assumptions
used in the model. The impact of the uncertainty of any
of the inputs in models can be assessed for robustness
using a sensitivity analysis, which accompanies most de-
cision analytic models. Finally, the user must assess
whether the time frame is appropriate for the disease
state, treatment pattern, and the environment to which
results are extrapolated.
Several decision analytic model analyses have been
conducted with an objective of determining if the higher
drug-acquisition costs of SSRIs and serotonin norepine-
phrine reuptake inhibitors (SNRIs) compared with TCAs
are offset by their superior tolerability profile (Frank
2001). The majority of these models have used the payer
perspective to test this hypothesis; a couple of models
have also taken a societal or semisocietal perspective by
including indirect costs, however. Regardless of the per-
spective, most of these modeling exercises have con-
cluded that SSRIs and/or SNRIs are a more cost-effective
alternative than TCAs.2 Building on this body of evi-
dence, the Canadian Coordinating Office for Health
Technology considered possible sequences of anti-
depressant therapies, a technique allowed by modeling.
Researchers determined that initiation of therapy with
2 Hylan 1996; Bentkover 1995; LaPierre 1995; Nuitjen 1995; Re-
vicki 1995, 1997; Jonsson 1994; Hatziandreu 1994; Einarson
1995, 1997.
CO S T - E F F E C T I V E N E S S
Months without
Total costs depression QALYs
$1,383 7.78 0.61
$1,580 8.72 0.60
$8,963 14.92 1.31
$5,343 17.09 1.31
an SSRI and subsequently switching to a TCA in the
event of intolerance or poor response is a more cost-
effective sequence than the reversed sequence or a TCA
alone, indicating that the use of SSRIs as first-line agents
is warranted (CCOHTA 1997).
Modeling also has been used to evaluate the effect of
increasing the length of antidepressant therapy, consis-
tent with national guidelines. In a decision analysis con-
ducted by Nuijten (2001), patients were initiated on SSRI
therapy and subsequently went into remission (while
remaining on an SSRI, switching to a TCA, or discon-
tinuing antidepressant therapy), were switched to a TCA,
or were hospitalized for severe depression. Outcome
probabilities were derived from the literature. Clinical
endpoints were time without depression (TWD) from
the third-party payer perspective and quality-adjusted life
years (QALYs) from the societal perspective. The eco-
nomic endpoints were direct medical costs and costs re-
sulting from lost productivity, which was based on num-
ber of lost working days. At 9 months, patients who went
into remission and discontinued antidepressant therapy
had substantially more TWD and no increase in the
number of QALYs. In addition, this group incurred fewer
costs and hence dominated (less costly and more effec-
tive) all other treatment patterns. When maintenance
treatment with antidepressant therapy was continued
for 21 months, costs decreased, TWD increased, and
QALYs did not change. Therefore, continuation of ther-
apy extending beyond 9 months to a minimum of 21
months, in accordance with guidelines, was determined
to be cost effective (Table 3).
The results described in Table 3 stem from decision
analytic models designed to simulate head-to-head com-
parisons between antidepressant therapies in a natural-
istic setting. Although the structured process and flexi-
bility make decision analytic modeling a valuable tool for
estimating the cost-effectiveness of various therapies,
their representation of actual clinical practice is only as
good as the assumptions and inputs used to populate the
DEPRESSION 37
 CO S T - E F F E C T I V E N E S S
model. Because most clinical inputs
used in these cost-effectiveness mod- FIGURE 1
els are based on clinical trial data, a
significant role has emerged for retro-
spective database analysis to assess the
economic effect of antidepressants in
a real-world environment.
RETROSPECTIVE
DATABASE ANALYSES
In an effort to overcome the exter-
nal validity quandary of the afore-
$5,610
mentioned study designs, retrospec-
tive database analyses draw upon
population-based information ob-
tained from large claims databases
(federal, state, or private insurers) to
provide real-world data. This study
design affords large sample sizes, the
targeting of a highly specific popula-
SOURCES:THOMPSON 1996, EADDY 2004
tion, and the capture of both resource
utilization and patient characteristics.
Although this method provides valuable information
regarding utilization in a naturalistic setting, it does not
offer the validity and bias protection afforded by RCTs.
Despite its nonexperimental design and limited internal
validity, retrospective database analyses offer insight into
actual clinical practice and its associated resource uti-
lization and cost.
Several retrospective database analyses have estab-
lished that patients treated with SSRIs are more likely to
be treated at an appropriate dose and duration than
those treated with TCAs.3 It was hypothesized that in-
appropriate doses and durations of therapy lead to
poorer clinical outcomes with an economic impact. Mc-
Combs (1990) identified and quantified the economic
consequences of subtherapeutic treatment in a retro-
spective database analysis. Patients who were not treated
with a TCA at a minimum therapeutic dose for 6 months
or more incurred an additional $1,000 per patient in the
first year of depression treatment as compared with those
patients treated at the clinically recommended dose and
duration. In addition, a series of retrospective database
analyses in network-model HMOs have established that
patients prescribed an SSRI, irrespective of dose, incur
significantly fewer health care expenditures relative to
those prescribed a TCA (P≤.05). 4 While the substantial economic implications of
Evidence-based guidelines recommend pharmaco-
therapy lasting a minimum of 90 days for acute man-
agement of depression and extension to a minimum of
3 Katzelnick 1996; Rosholm 1993; Sclar 1994, 1995, 1997, 1998;
Skaer 1995, 1996.
4 Sclar 1994, 1995, 1997; Skaer 1995, 1996.
38 P&T DIGEST
Costs accrued by pattern of selective serotonin
reuptake inhibitor use: results from two analyses
Early discontinuation
Switching/augmentation
Upward titration
Partial compliance
Continuation therapy
$7,590 $7,858
$6,289 $6,375
$5,909
$4,479 $5,143
$3,822
$3,393
Thompson Eaddy
180 days for continuation therapy to achieve optimal
outcomes (APA 2000). Thompson (1996) assessed the
economic consequences of compliance with these rec-
ommendations in a retrospective database analysis of
patients initiating therapy for depression with an SSRI.
This analysis classified patients into one of five patterns
of antidepressant-use cohorts: 1) early discontinuation,
2) switching/augmentation, 3) upward titration, 4) par-
tial compliance, and 5) 3-month use (Figure 1). Patients
who switched therapies or augmented their pharmaco-
therapy with another agent incurred the highest total
medical costs ($7,590). The lowest costs were observed
in patients continuing their use of antidepressant ther-
apy for a minimum of 3 months ($3,393). Similar results
were seen in a recent study conducted by Eaddy (2004),
which used an analysis comparable to Thompson’s to de-
termine if the relationship between length of therapy
and health care expenditures is similar in the current
managed care environment. As with Thompson’s find-
ings, patients changing or augmenting antidepressant
pharmacotherapy incurred the highest total medical
costs ($7,858). The lowest total medical costs were in-
curred by patients remaining on therapy for at least 90
days ($5,143) (Figure 1).
switching antidepressant pharmacotherapies have been
defined, clinical predictors of switching agents also have
been assessed in retrospective database analyses. Sclar
(1998) showed that patients initiating therapy with a
TCA were 3 to 4 times more likely to switch agents com-
pared with patients initiating therapy with an SSRI. As

expected, patients in the TCA cohort accrued higher
heath care expenditures than patients in the SSRI cohort.
Russell (1999) conducted a similar analysis comparing
three SSRIs: fluoxetine, immediate-release paroxetine,
and sertraline. Patients beginning therapy with fluoxe-
tine were the least likely to switch antidepressant agents
(P=.001) compared with patients initiating immediate-
release paroxetine or sertraline therapy. In contrast to
previous results, total health care expenditures did not
differ significantly between the groups even though the
rate of switching within the treatment groups differed.
As evidenced by Thompson (1996) and Eaddy (2004),
increasing length of antidepressant therapy to a duration
consistent with national clinical guidelines translates to
a reduction in total health care expenditures. With that
premise in mind, researchers investigated whether there
were differences in length of therapy based on the indi-
vidual antidepressant that was prescribed for new anti-
depressant cases (APA 2000, Thompson 1996, Eaddy
2004). Russell (1999) and Crown (2001) compared the
durations of therapy for patients initiating therapy with
sertraline, immediate-release paroxetine, and fluoxetine
and examined the costs incurred by patients in each
group. In both analyses, patients initiated on fluoxetine
were more likely to achieve treatment durations consis-
tent with recommendations (Russell P<.001; Crown
P<.05) as compared with patients initiated on sertraline
or immediate-release paroxetine. Significant differences
in costs were not observed between patients
in the three groups. In a similar analysis, FIGURE 2 Discontinuation rates for controlled-
Polsky (2002) determined that patients ini-
tiated on therapy with fluoxetine experi-
enced an interruption in treatment (defined
as a 30-day lapse in prescriptions) signifi-
cantly later than patients initiated on ser-
Patients remaining on therapy (%)
traline or immediate-release paroxetine. In-
terruption of therapy was not shown to be
related to health care costs.
Eaddy (2003) expanded on these analyses
by using survival analysis to compare the
time to discontinuation for each of the avail-
able SSRIs on the market at the time of their
analysis. Patients who received one of the
immediate-release SSRIs or controlled-
release paroxetine between April 1, 2002,
and Dec. 31, 2002, were identified in a large
commercial managed care database. Patients
were required to have 6 months of enroll-
ment data before their index date, defined as
the date of the first SSRI prescription, with-
out evidence of antidepressant therapy. Only
CR=controlled release; IR=immediate release.
patients experiencing new therapy with
SOURCE: SHEEHAN 2004
SSRIs were included in the study.
CO S T - E F F E C T I V E N E S S
All prescriptions were given a starting date (when the
prescription was filled) and ending date (corresponding
to the starting date of the prescription plus the number
of days for which medication was supplied). Patients
were deemed to have discontinued therapy when more
than 15 days elapsed between prescriptions. A patient’s
time to discontinuation was calculated from the index
date to the ending date of the last prescription prior to
the 15-day gap. Patients receiving controlled-release
paroxetine were less likely to discontinue therapy when
compared with those receiving each immediate-release
SSRI (P<.0001). Patients who received fluoxetine were
least likely of all those receiving immediate-release SSRIs
to discontinue therapy, while patients receiving imme-
diate-release paroxetine were most likely to discontinue.
Building on these findings, Sheehan (2004) investi-
gated the economic consequences associated with early
discontinuation by comparing patients on immediate-
release paroxetine with patients receiving controlled-
release paroxetine. A total of 3,500 patients initiating
therapy with paroxetine immediate- or controlled-release
were included in this analysis and stratified into two
groups, those with and without a diagnosis for an ap-
proved indication for their study agent. After controlling
for age, gender, mental health diagnoses, presence of
mental health specialty care, and Charlson comorbidity
scores, the risk of early treatment discontinuation was as-
sessed. Controlled-release paroxetine reduced the risk of
and immediate-release paroxetine
100
Paroxetine CR
90 Paroxetine IR
80
70
60
50
40
30
0 30 60 90 120 150 180
Days of therapy
DEPRESSION 39
 CO S T - E F F E C T I V E N E S S
TABLE 4 Expenditures for immediate-release
versus controlled-release paroxetine
paroxetine CR
Indication-specific analysis ($)
Monthly medical charges*
Monthly antidepressant charges†
Total
Difference
Monthly medical charges‡
Monthly antidepressant charges†
Total
Difference
Controlling for age, gender, diagnostic information, and Charlson comorbidity index scores.
*P=.054, †P<.0001, ‡P=.1018.
SOURCE: SHEEHAN 2004
early treatment discontinuation by 40 percent in the
indication-specific analysis (Figure 2, page 39) and by 35
percent in the analysis independent of indication.
An examination of expenditures in this analysis re-
vealed that patients initiating therapy with paroxetine
immediate-release incurred significantly higher costs
than those initiating therapy with paroxetine controlled-
release (Table 4) in both the indication-specific and non-
indication analyses. Antidepressant-related pharmacy
expenditures were higher in the controlled-release parox-
etine cohort; differences in medical costs offset this in-
crease by $59 to $109, however.
Discontinuation of antidepressant pharmacotherapy
before the recommended length of therapy has been
shown to be associated with an increased risk of relapse
or recurrence in several retrospective database analyses
(Sood 2000, Melfi 1998, Croghan 1998). In addition, pa-
tients remaining on their index SSRI for at least 120 days
at a consistent dose were shown to be least likely to ex-
perience a relapse or recurrence relative to other pattern
of use groups (Claxton 2000).
CONCLUSIONS
Five pharmacoeconomic study designs have been pre-
sented, inherent strengths and weaknesses considered,
and depression-specific examples have been provided.
Overall, three general conclusions may be drawn from the
results presented herein:
• RCTs and prospective naturalistic inquiries have
shown newer and older antidepressant agents to be
40 P&T DIGEST
approximately equal in terms
of efficacy. A series of decision
analytic models have employed
these clinical data and shown
paroxetine IR newer agents (SSRIs and
SNRIs) to be cost-effective rel-
ative to older agents.
414 532 • Cost-effectiveness data regard-
77 67
ing the individual newer anti-
491 599
depressants are limited and not
109 as clear. Nonetheless, there is a
growing body of literature de-
Nonindication analysis ($) fining the clinical and econo-
519 589 mic differences between these
77 66 agents.
596 655 • Multiple retrospective database
59 analyses have shown an eco-
nomic benefit associated with
adherence to clinical guidelines
regarding the length of anti-
depressant therapy.
As evidence-based medicine evolves, the importance
of pharmacoeconomic data increases. The prudent prac-
titioner will strive to understand pharmacoeconomic
methodologies, each study design’s role in the pharma-
ceutical product’s life cycle, and the relevant literature re-
lating to the specific area of practice. The knowledge
gained through the study of pharmacoeconomics is piv-
otal in formulary decision making.
REFERENCES
Anderson IM, Tomenson BM. Treatment discontinuation with
selective serotonin reuptake inhibitors compared with tri-
cyclic antidepressants: a meta-analysis. Br Med J.
1995;310:1433–1438.
APA (American Psychiatric Association). Practice guideline for
the treatment of patients with major depressive disorder
(revision). Am J Psychiatry. 2000;157(4 suppl):1–45.
Bentkover JD, Feighner JP. Cost analysis of paroxetine versus
imipramine in major depression. Pharmacoeconomics.
1995;8:223–232.
Boyer P, Danion JM, Bisserbe JC, et al. Clinical and economic
comparison of sertraline and fluoxetine in the treatment of
depression: a 6-month, double-blind study in a primary
care setting in France. Pharmacoeconomics. 1998;13(1 pt
2):157–169.
CCOHTA (Canadian Coordinating Office for Health Technol-
ogy Assessment). Selective serotonin reuptake inhibitors
(SSRIs) for major depression. Part 2. The cost effectiveness
of SSRIs in the treatment of depression. Ottawa: Canadian
Coordinating Office for Health Technology Assessment.
1997.
Claxton AJ, Li Z, McKendrick J. Selective serotonin reuptake in-
hibitor treatment in the UK: risk of relapse or recurrence of
depression. Br J Psychiatry. 2000;177:163–168.
Croghan TW, Melfi CA, Crown WE, Chawla A. Cost effective-
ness of antidepressant medications. J Mental Health Policy
Econ. 1998;1(3):109–117.

Crown WH, Treglia M, Meneades L, White A. Long-term costs
of treatment for depression: impact of drug selection and
guideline adherence. Value in Health. 2001;4(4):295–307.
Eaddy M, Bramely T, Regan T. Time to antidepressant discontin-
uation: a comparison of controlled-release paroxetine and
immediate-release selective serotonin reuptake inhibitors.
Manag Care Interface. 2003;16(12):22–27.
Eaddy M, Sarnes M, Regan T, Mauch R. SSRI utilization patterns
and the economic consequences of early discontinuation.
[Poster]. Presented at: American Psychiatric Association
157th Annual Meeting, May 1–6, 2004. New York.
Einarson TR, Addis A, Iskedjian M. Pharmacoeconomic analysis
of venlafaxine in the treatment of major depressive disor-
der. Pharmacoeconomics. 1997;12(2 pt 2):286–296.
Einarson TR, Arikian S, Sweeney S, Doyle J. A model to evaluate
the cost effectiveness of oral therapies in the management
of patients with major depressive disorders. Clin Ther.
1995;17:136–153.
Frank L, Revicki DA, Sorensen SV, Shih YC. The economics of
selective serotonin reuptake inhibitors in depression: a crit-
ical review. CNS Drugs. 2001;15(1):59–83.
Freemantle N, Mason J. The importance of achieving additional
drug benefits at a reasonable cost: a review of the fluoxetine
years. Pharmacoeconomics. 2000;17(4):319–324.
Gallo JJ. TCAs vs. SSRIs. Same bang for whose buck? Arch Fam
Med. 1999;8:326–327.
Goff VV. Depression: a decade of progress, more to do. Wash-
ington: National Health Policy Forum Issue Brief No. 786.
2002.
Hatziandreu EJ, Brown RE, Revicki DA, et al. Cost utility of
maintenance treatment of recurrent depression with ser-
traline versus episodic treatment with dothiepin.
Pharmacoeconomics. 1994;5(3):249–268.
Hosak L, Tuma I, Hanus H, Straka L. Costs and outcomes of use
of amitriptyline, citalopram, and fluoxetine in major de-
pression: exploratory study. Acta Medica. 2000;43(4):
133–137.
Hotopf M, Lewis G, Normand C. Are SSRIs a cost-effective alter-
native to tricyclics? Br J Psychiatry. 1996;168:404–409.
Hylan TR, Kotsanos JG, Andersen JS, et al Comparison of a de-
cision analytic model with results from a naturalistic eco-
nomic clinical trial: an application to evaluating alternative
antidepressants. Am J Manag Care. 1996;2:1211–1223.
Jonsson B, Bebbington PE. What price depression? The cost of
depression and the cost–effectiveness of pharmacological
treatment. Br J Psychiatry. 1994;164:665–673.
Katzelnick DJ, Koback KA, Jefferson JW, et al. Prescribing pat-
terns of antidepressant medications for depression in a
HMO. Formulary. 1996;31:374–388.
LaPierre Y, Bentkover J, Schainbaum S, Manners S. Direct cost of
depression: analysis of treatment costs of paroxetine versus
imipramine in Canada. Can J Psychiatry. 1995;40:370–377.
McCombs JS, Nichol MB, Stimmel GE, et al. The cost of antide-
pressant drug therapy failure: a study of antidepressant use
patterns in a Medicaid population. J Clin Psychiatry.
1990;51(suppl 6):60–69.
Melfi CA, Chawla AJ, Croghan TW, et al. The effects of adher-
ence to antidepressant treatment guidelines on relapse and
recurrence of depression. Arch Gen Psychiatry. 1998;55:
1128–1132.
Montgomery SA, Henry J, McDonald G, et al. Selective sero-
tonin reuptake inhibitors: meta-analysis of discontinuation
rates. Int Clin Psychopharmacol. 1994;9:47–53.
Montgomery SA, Kasper S. Comparison of compliance between
serotonin reuptake inhibitors and tricyclic antidepressants:
A meta-analysis. Int Clin Psychopharmacol. 1995;9(suppl
4):33–40.
CO S T - E F F E C T I V E N E S S
Nuijten MJC, Hardens M, Souetre E. A Markov process analysis
comparing the cost–effectiveness of maintenance therapy
with citalopram versus standard therapy in major depres-
sion. Pharmacoeconomics. 1995;8:159–168.
Nuijten MJC. Assessment of clinical guidelines for continuation
treatment in major depression. Value in Health. 2001;4:
281–294.
Polsky D, Onesirosan P, Bauer MS, Glick HA. Duration of ther-
apy and health care costs of fluoxetine, paroxetine, and ser-
traline in 6 health plans. J Clin Psychiatry. 2002;63:156–164.
Revicki DA, Brown RE, Keller MB, et al. Cost effectiveness of
newer antidepressants compared with tricyclic anti-
depressants in managed care settings. J Clin Psychiatry.
1997;58(2):47–58.
Revicki DA, Brown RE, Palmer W, et al. Modeling the cost-
effectiveness of antidepressant treatment in primary care.
Pharmacoeconomics. 1995;8(6):524–540.
Rosholm JU, Hallas J, Gram LF. Outpatient utilization of anti-
depressants: a prescription database analysis. J Affect Disord.
1993;27:21–28.
Russell JM, Berndt ER, Miceli R, et al. Course and cost of treat-
ment for depression with fluoxetine, paroxetine, and sertra-
line. Am J Manag Care. 1999;5:597–606.
Sclar DA, Robison LM, Skaer TL, et al. Antidepressant pharma-
cotherapy: economic outcomes in a health maintenance or-
ganization. Clin Ther. 1994;16:715–730.
Sclar DA, Roison LM, Skaer TL, et al. Antidepressant pharma-
cotherapy: economic evaluation of fluoxetine, paroxetine,
and sertraline in a health maintenance organization. J Int
Med Res. 1995;23:395–412.
Sclar DA, Skaer TL, Robison LM, et al. Evidence-based algorithm
for antidepressant pharmacotherapy. JAMA Southeast Asia.
1997;13(Suppl 1):33–40.
Sclar DA, Skaer TL, Robison LM, et al. Economic outcomes with
antidepressant pharmacotherapy: a retrospective intent-to-
treat analysis. J Clin Psychiatry. 1998;59(suppl 2):13–17.
Sheehan D, Sarnes M, Regan T, Eaddy M. The economic conse-
quences of treatment discontinuation: paroxetine con-
trolled release versus paroxetine immediate release.
[Poster]. Presented at: American Psychiatric Association
157th Annual Meeting, May 1–6, 2004. New York.
Simon GE, Heiligenstein J, Revicki D, et al. Long-term outcomes
of initial antidepressant drug choice in a real-world ran-
domized trial. Arch Fam Med. 1999;8:319–325.
Simon GE, VonKorff M, Heilgenstein JH, et al. Initial antidepres-
sant choice in primary care: effectiveness and cost of
fluoxetine vs. tricyclic antidepressants. JAMA. 1996;275:
1897–1902.
Skaer TL, Sclar DA, Robison LM, et al. Economic valuation of
amitriptyline, desipramine, nortriptyline, and sertraline in
the management of patients with depression. Curr Ther Res.
1995;56:556–567.
Skaer TL, Sclar DA, Robison LM, et al. Antidepressant
pharmacotherapy: effect of women’s resource utilization
within a health maintenance organization. J Appl Ther.
1996;1:45–52.
Skaer TL, Sclar DA, Robison LM, Galin RS. The need for an iter-
ative process for assessing economic outcomes associated
with SSRIs. Pharmacoeconomics. 2000;18;205–214.
Sood N, Treglia M, Obenchain RL, et al. Determinants of anti-
depressant treatment outcome. Am J Manag Care. 2000;6:
1327–1336.
Thompson D, Buesching D, Gregor KJ, Oster G. Patterns of anti-
depressant use and their relation to costs of care. Am J
Manag Care. 1996;2:1239–1246.
DEPRESSION 41
 CO M P L I A N C E CO M P L I A N C E
Adherence With Antidepressant
Medication
VINCENT J. GIANNETTI, PHD1
Duquesne University
SUMMARY
Nonadherence with antidepressant medi-
cations is a well-documented factor in treat-
ment failure.This article uses the literature
to build a collaborative counseling model
to increase antidepressant adherence. Stud-
ies may be warranted to determine whether
specific antidepressants, when used as part
of a collaborative strategy, can improve
long-term adherence rates.
N
onadherence can be defined as deviating from
a prescribed course of treatment and can in-
clude not filling a prescription, premature dis-
continuance of a regimen, dosing errors,
and/or self-initiated changes. Up to 20 percent of patients
do not fill an initial prescription (Burns 1992). Estimates
of noncompliance for all medical conditions range from
20 to 80 percent of the population, leading to subthera-
peutic benefit (Dunbar-Jacob 1995) and costing more
than $100 billion annually, because of increased medical
and other costs (Grahl 1994).
Antidepressant medication is a critical component in
the effective treatment of patients with depression. Clin-
ical guidelines for the treatment of depression recom-
mend the continuation of antidepressant medication
for 4 to 9 months after the resolution of depressive symp-
toms (AHCPR 1993, Schulberg 1998). Nevertheless, pre-
1 Vincent Giannetti, PhD, is professor of social and ad-
ministrative sciences in pharmacy at the Duquesne Uni-
versity School of Pharmacy, where he teaches in the psy-
chiatry section of the therapeutics course. He also teaches
organizational behavior and health care management in
the School of Business. Giannetti is a licensed psychologist
in Pennsylvania and has graduate degrees in psychology,
public health, and social work. His publications have fo-
cused on health care ethics, substance abuse, and behavioral
health. He maintains a private psychotherapy practice in
Pittsburgh.
42 P&T DIGEST
mature discontinuation can
reach as high as 44 percent
as soon as 3 months after
initiation of antidepressant
treatment (Katon 1995). By
contrast, the average non-
adherence rate across con-
ditions is nearly 25 percent
(DiMatteo 2004).
Adherence interventions
have occurred in a context
of lacking clear specific
guidelines that are gener-
ated by scientifically vali-
dated studies that are de- Vincent J. Giannetti, PhD
signed to show which
approaches are most effective in ensuring adherence. A
number of theoretical approaches to adherence are dis-
cussed in the literature, including the biomedical model,
social learning theory, operant conditioning, rational
belief approaches, and communication and self-regula-
tion theory (Leventhal 1987). Although these models
have utility for researchers, they often are difficult to
convert into concrete, brief interventions that practi-
tioners can use. This article reviews the adherence liter-
ature and looks at how it relates specifically to anti-
depressant medication, and suggests brief interventions
that have empirical support and can be effective for in-
creasing adherence with drug therapies for depression.
ADHERENCE COUNSELING MODEL
The adherence counseling model developed herein
integrates counseling theory with specific interventions
that have some research support for facilitating adher-
ence. The adherence model does not address cost and ac-
cess to medication, although these can be barriers to ad-
herence. These policy issues need to be addressed on a
health care system level. Individual practitioners at least
should assess inability to pay for medications as a po-
tential barrier, however, and should attempt to resolve
these issues by referral to appropriate human service
agencies. The Pharmaceutical Research and Manufac-

turers of America has an assistance program for unin-
sured and low-income patients that can be assessed on-
line at «www.needymeds.com». Also, this model assumes
accurate diagnosis, specification of target symptoms for
treatment, and tracking of symptom relief throughout
the course of antidepressant treatment. The establish-
ment of a baseline measure of symptoms using stan-
dardized scales (Sajatovic 2001) and periodic evalua-
tion of amelioration of symptoms after initiation of
antidepressant medication is assumed to be a part of
good practice in the treatment of depression and a min-
imum intervention tool to assure adherence and thera-
peutic outcomes.
Table 1 summarizes the adherence counseling model.
Utilized in a collaborative approach to health care, this
model represents the basic elements of counseling for ad-
herence with antidepressant medication.
The therapeutic alliance
The culture of health care delivery has evolved from
an active-passive model to a shared decision-making
style preferred by most patients and leading to patients’
satisfaction with their health care (Bertakis 1991). These
philosophical shifts in health care delivery have been
paralleled by activism among consumers. Many con-
sumers feel empowered, owing to an explosion in the
amount of health care information that has become
available from varied sources. This change makes it all the
more important for clinicians to hone their communi-
cation skills, especially if miscommunication fosters
nonadherence.
The basics of good communication skills include in-
volving the patient, addressing patient concerns, offer-
ing options and explanations, and periodically checking
for understanding. Involvement of the patient should
also include eliciting the patient’s perspective, which ex-
tends to determining how much the patient desires to
participate in decision making; assessing the patient’s
perception of the disease and its treatment; and en-
couraging the patient to express concerns about regi-
mens. Treatment options should be discussed and
information tailored to the concerns of the patient
(Maguire 2002).
This assessment and information sharing occurs
within the context of a professional relationship. The pri-
macy of relationship has become less of a focus in the
time-constrained “business” of medicine. Encouraging
active engagement in the therapeutic process has been a
hallmark of psychotherapeutic treatment but has been
lost to an occasional overemphasis on technological ad-
vances in medicine. Too often, practitioners view med-
ication from their perspective of specialized medical
training and fail to understand the patient’s unique ap-
CO M P L I A N C E
TABLE 1 Adherence counseling model
1. Establish therapeutic alliance with patient.
2. Destigmatize the illness.
3. Encourage and explain positive health beliefs
regarding depression.
4. Encourage and explain the positive perception
of benefits outweighing barriers regarding
medication taking.
5. Dispel and dispute irrational health beliefs
regarding disease state and medication.
6. Use behavioral interventions to tailor the regimen
or choice of medicine and dosage form to the
lifestyle of the patient and reduce the impact
of side effects.
7. Use collaborative agreements and relationships
with pharmacists and psychotherapists to
continually monitor and evaluate adherence.
propriation of subjective meaning to the medication-tak-
ing process (Conrad 1985). In its simplest form, adher-
ence counseling involves understanding the patient’s
unique and sometimes idiosyncratic perspective and
then attempting to modify the perspective to align beliefs,
feelings, and behaviors more favorably with effective
medication-taking behavior. Problems related to effective
communication of information to patients, patients’ un-
derstanding of information, and the lack of collaboration
between practitioner and patient have been found to
lead to poor adherence (DiMatteo 1994). Within this
context, an essential first step in the adherence counsel-
ing process is the demonstration of active empathy, non-
judgmental acceptance, and genuine understanding of
the illness from the patient’s perspective. Numerous stud-
ies concerning positive behavioral change in counseling
have documented the critical importance of the presence
of these practitioner qualities (Krupnick 1996).
Health beliefs
Nonadherence to drug regimens has been related to
patients feeling stigmatized by their illness and experi-
encing threats to their self-reliance (Conrad 1985). The
first step in adherence counseling is to explain the neuro-
biological basis of depression, so that the patient can
perceive depression as a medical condition and not a
function of weak character or a flawed personality.
Medication-taking can then be reframed as an action by
the patient that enables him or her to take control of the
illness by compensating for the biochemical deficit in
naturally occurring mood-altering substances. Then spe-
cific health beliefs can be addressed.
The health belief model has been demonstrated to be
useful in facilitating medication adherence. The patient’s
DEPRESSION 43
 CO M P L I A N C E
cognitive representation of the illness may be at variance
with the provider’s perspective and play a significant
role in nonadherence (Scott 2002). The health belief
model hypothesizes that the probability that a patient will
adhere to a therapeutic regimen is a function of the per-
ceived susceptibility to negative consequences of an ill-
ness, as well as being a function of the perceived value of
adherence to a regimen versus barriers and costs (in-
cluding side effects, inconvenience, and the stigma asso-
ciated with an illness) (Becker 1975). The model argues
that adherent patients will accept fully that they have the
illness of depression, will understand the negative emo-
tional and physical consequences, and will recognize the
value of the benefits of antidepressant medications rel-
ative to any barriers. In practice, this means the patient
must understand the illness and how the medication
compensates for the illness, and the patient must make
a rational calculus that the benefits of taking the medi-
cation outweigh any inconveniences or costs.
Once the stigma of the illness has been resolved and
an understanding of the illness communicated, specific
information regarding side effects and patient measures
to control or ameliorate them should be the focus of
counseling. Some physicians selectively provide patients
with information about the side effects that might be en-
countered with a selective serotonin reuptake inhibitor
(SSRI) or a serotonin norepinephrine reuptake inhibitor
(SNRI), believing that imparting the knowledge itself
may engender reports of side effects, real or not, from pa-
tients. One study, however, showed that 55 percent of pa-
tients receiving an SSRI may experience at least one ad-
verse side effect described as “a lot” or “extremely”
bothersome (Bull 2002). In the interest of increasing the
likelihood that patients will adhere to therapy, a complete
discussion of a drug’s side-effect profile may be war-
ranted when treatment options are presented. Such a
discussion should encompass not only the potential side
effects, but also strategies for managing them. In coun-
seling patients about potential side effects of an anti-
depressant medication, the practitioner should also de-
termine whether the patient already experiences these
problems so they are not erroneously ascribed to the
medication (Rollman 1997). Along with patients’ favor-
able attitudes toward antidepressant therapy, patients’
confidence in managing the side effects of drug therapy
is a statistically significant predictor of adherence to
long-term (up to 12 months) therapy (Lin 2003).
Both encouraging the patient’s belief in the value of the
regimen and addressing the patient’s concerns about the
therapy will enhance adherence (DiMatteo 1993). Finally,
the choice of antidepressant should involve a side-effect
profile that addresses the needs, desires, and lifestyle of
the patient.
44 P&T DIGEST
TABLE 2 Patient beliefs that contribute to
nonadherence
Medication is addicting.
Medication is a crutch.
Medication is only needed occasionally when I am
really depressed.
I will be unable to tolerate the side effects.
I will never be able to discontinue the medication.
If I don’t feel well immediately, the medication isn’t
working.
The medicine cannot solve my problems.
Medication will make me tired all the time.
The second stage in addressing health beliefs is to as-
sess and then educate patients regarding dysfunctional
perceptions of medication. These underlying beliefs will
vary for each patient. Many of these dysfunctional beliefs
for antidepressant medication have been catalogued
(Beck 1979). They are summarized in Table 2.
Many of these beliefs can be addressed after soliciting
the patient’s feelings regarding the taking of medication.
For example, the lag time between the initiation of an
SSRI or SNRI and therapeutic effect should be addressed
routinely, because patients expecting immediate symp-
tomatic relief may conclude mistakenly that the med-
ication is ineffective. Other patient beliefs should be ad-
dressed in a respectful but matter-of-fact discussion
about the medication. For example, the belief that med-
ication is a crutch can be reframed by explaining that
concentration deficits and low energy are core symptoms
of depression and can be helped by medication, thus
freeing the patient to take the measures necessary to re-
cover. Medication can be explained as offering enough
“biological comfort and freedom from symptoms” to
help the patient make changes that he or she feels are nec-
essary. The existence of these negative beliefs and the spe-
cific refutation of these beliefs will vary with each patient.
Cueing and monitoring
Once positive beliefs have been encouraged and nega-
tive beliefs are brought to light, the mechanics of adher-
ence can be addressed. A simple explanation of the dos-
ing regimen will not ensure adherence, because
forgetfulness can be a contributory factor in nonadher-
ence. For antidepressant medication, according to one
study, 72 percent of physicians reported telling patients
to take an SSRI for at least 6 months, while only 34 per-
cent of the patients recalled receiving that information.
The authors stated that patients’ forgetfulness could ac-
count for at least part of the communication gap, thus
emphasizing the need for periodic reinforcement of in-
structions (Bull 2002). It is worthy of note, however, that

in relying on their own memories, physicians may also
have overstated the rate at which they provided patients
with information.
The most effective behavioral techniques to ensure ad-
herence are cueing and monitoring. Attaining stimulus
control through cueing can be accomplished by pairing
a dosing regimen with naturally occurring rituals in the
patient’s life. Establishing recurrent cues for medication
taking in the patient’s environment has been found to be
associated with increased adherence (Haynes 1976,
Logan 1979). Associating medication taking with meals
and other routine activities, as well as enlisting family
members for periodic reminders or using medication or-
ganizers with slots for days and times, can be simple
low-cost techniques to cue memory.
Monitoring and reinforcing medication adherence
once medication is initiated can be especially effective.
It has been shown that physician and pharmacist moni-
toring of patient response increased both patient satis-
faction and adherence with antidepressant medication
(Bull 2002, Bultman 2002). Physician follow-up of three
or more visits increased adherence, and pharmacist
problem-solving and support with medication prob-
lems after initiation of therapy were associated with bet-
ter adherence for first-time users. Follow-up monitoring
reminds the patient of the regimen’s importance, con-
firms the practitioner’s concern for the patient, and re-
inforces the importance of adherence in the patient’s
mind through the practitioner’s investment of time and
attention. Monitoring also can address the occurrence of
side effects.
TEAM APPROACH
Treatment of depression is frequently delivered in
primary care settings. The combination of medication
and psychotherapy represents an effective approach to
treating depression (Craighead 1998). Behavioral health
is now more frequently integrated into health care, and
a number of brief-oriented methods of psychotherapy
have been found to be effective with depression (Wells
1990); more recently, Schulberg (2002) found that
though psychotherapy is more costly than a primary
care physician’s usual care, a depression-specific psycho-
therapy produces better clinical outcomes than a primary
care physician’s usual care and outcomes similar to those
produced by pharmacotherapy.
Prescribing physicians can utilize psychotherapists
through behavioral health networks to monitor thera-
peutic response and discuss and reinforce adherence,
paying particular attention to such psychological barri-
ers as guilt and demoralization as part of the psycho-
therapeutic enterprise. Pharmacists can play an impor-
tant role in identifying early stages of nonadherence and
CO M P L I A N C E
also are educated to provide pharmaceutical care by
identifying and responding to drug-related problems
(Bultman 2002). Within primary care practices, depres-
sion care managers have been found highly effective in
supporting patient adherence to antidepressant medi-
cations and monitoring side effects (Unützer 2002, Bruce
2004).
These professionals can assist physicians in busy prac-
tices to support adherence. This is especially true be-
cause treatment for depression can be chronic, and in-
formation is better retained when spread out over
multiple sessions rather than amassed in one session.
Also, patients’ needs and conditions change over time,
thus reinforcing the need for continued monitoring.
Managed care organizations are well positioned to en-
courage collaboration and case management within their
networks.
Within a managed care setting, the use of clinical
pharmacy specialists to augment psychiatric services
provided for patients with mild to moderate depression
has resulted in a higher medication-possession ratio
(0.81 vs. 0.66), higher switching rates (24 vs. 5 percent),
and a reduction in subsequent office visits to primary
care physicians (39 vs. 12 percent), compared with a
control group during a 6-month follow-up period (Fin-
ley 2002).
On the other hand, there are limits to what MCOs can
accomplish with labor-intensive interventions, given the
competitive environment in which MCOs function. This
makes it of paramount importance for MCOs to exploit
any characteristics of antidepressant medications in and
of themselves that might improve adherence.
From this perspective, a recent study of discontinua-
tion rates for SSRIs merits examination, especially be-
cause it was based on real-world data from managed
care environments. This retrospective analysis of a large
managed care database (36 million lives covered by more
than 60 MCOs) showed that, after 180 days, patients
who received a controlled-release SSRI (paroxetine) were
28 percent less likely to discontinue treatment than pa-
tients receiving an immediate-release SSRI (Eaddy 2003).
The study examined medical and pharmacy claims
for more than 82,000 patients who experienced new
SSRI therapy from April 2002 through December 2002
for a depressive disorder or an anxiety disorder, or both.
Patients lacking a diagnosis of depression or anxiety
were classified as undiagnosed and were included in the
study, but patients diagnosed with schizophrenia or a
bipolar disorder were excluded, as were patients who
were receiving antipsychotics.
In this population, sertraline was prescribed most fre-
quently (28 percent), followed by citalopram (25 per-
cent), immediate-release paroxetine (23 percent), fluoxe-
DEPRESSION 45
 CO M P L I A N C E
FIGURE SSRI persistence
100
Patients remaining on therapy (%)
90
80
70
60
50
40
30
0 20 40 60
Survival curves for selective serotonin reuptake inhibitors.
CR=controlled release; IR=immediate release. All differences between the CR form and
IR forms at 180 days are significant at P<.0001.
SOURCE: EADDY 2003
tine (16 percent), and controlled-release paroxetine (8
percent). The incidence of anxiety-related diagnoses was
higher among the patients receiving controlled-release
paroxetine than among patients receiving an immediate-
release SSRI (32.5 vs. 19.3 percent). After 30 days, about
75 percent of patients remained on their therapy, regard-
less of what it was; after 60 days, about 70 percent; and after
90 days, about 60 percent. At each of these time points, the
percentages of patients remaining on controlled-release
paroxetine and fluoxetine were virtually identical, with the
rates for the other immediate-release SSRIs being slightly
lower (and immediate-release paroxetine, lowest of all).
From that point on, however, the curves gradually sepa-
rated, and after 180 days, 55 percent of patients remained
on controlled-release paroxetine versus 43 percent on the
immediate-release SSRIs (Figure). Adjustments for de-
mographic and clinical variables showed a 28 percent re-
duction in the risk of treatment discontinuation among
patients receiving controlled-release paroxetine versus an
immediate-release SSRI. The study was not designed to
identify reasons for discontinuation.
These results gain added interest, however, when con-
sidered in the context of randomized controlled studies
demonstrating that certain interventions that were seen to
improve treatment adherence (and depressive symptoms)
at 4 and 7 months (Katon 1995, Katon 1996) did not lead
to better adherence or better outcomes than that achieved
with routine care after 19 months (Lin 1999). The en-
46 P&T DIGEST
hanced interventions, which involved
education of patients and physicians,
along with a reorganization of primary
care services, were provided during the
Paroxetine CR
first 12 weeks of therapy, after which pa-
Citalopram
tients reverted to routine care. These
Paroxetine IR
studies were conducted prior to the avail-
Fluoxetine
ability of controlled-release paroxetine.
Sertraline
Given the financial constraints that pre-
sumably would preclude extending en-
hanced interventions much beyond 12
weeks in many real-world settings, addi-
tional studies with extended-release
paroxetine, delayed-release fluoxetine,
and extended-release venlafaxine might
be warranted in the interests of ascer-
taining whether these products can im-
80 100 120 140 160 180 prove long-term outcomes.
Day
SUMMARY
Methods of counseling for adher-
ence lack a well-researched experi-
mental basis for determining which
approaches are the most effective. An
extensive history of descriptive adher-
ence research, however, can provide data for building a
multifaceted model of adherence counseling.
First, the health care provider must develop a caring
and collaborative relationship with patients for any coun-
seling to be effective. Second, positive beliefs and attitudes
concerning the regimen must be encouraged, and irra-
tional and negative beliefs and attitudes must be dis-
puted. Third, patients must view the benefits of the regi-
men as outweighing any barriers. Fourth, the regimen
must be tailored to the patient’s needs and the patient
must be given cognitive-behavioral control over the reg-
imen by anticipating, understanding, and managing side
effects. Fifth, simplification of the dosing regimen should
be considered through the possible use of controlled-
release dosage forms.
Finally, periodic monitoring and reinforcement using
pharmacists and psychotherapists in a collaborative
and comprehensive model of physical and psychologi-
cal care will offer the most effective approach to long-
term treatment.
REFERENCES
AHCPR. Depression in Primary Care: Vol. 2. Treatment of Major
Depression (Clinical Practice Guideline No. 5). Rockville,
Md.: AHCPR. 1993b.
Beck AT, Rush AJ, Shaw BF, Emery G. Cognitive Therapy of De-
pression. New York: Guilford Press. 1979.
Becker MH, Maiman LA. Sociobehavioral determinants of com-
pliance with health and medical care recommendations.
Med Care. 1975;13:10–24.

Bertakis KD, Roter D, Putnam SM. The relationship of physician
medical interview style to patient satisfaction. J Fam Pract.
1991;32:175–181.
Bruce ML, Ten Have TR,Reynolds CF III. Reducing Suicidal
Ideation and Depressive Symptoms in Depressed Older
Primary Care Patients: A Randomized Controlled Trial.
JAMA. 2004;291:1081-1091.
Bull SA, Hu XH, Hunkeler EM, et al. Discontinuation of the use
and switching of antidepressants: influence of patient-
physician communication. JAMA. 2002;288:1403–1409.
Bultman DC, Svarstad BL. Effects of pharmacist monitoring on
patient satisfaction with antidepressant medication. J Am
Pharm Assoc. 2002;42:36-43.
Burns JM, Sneddon I, Lovell M, et al. Elderly patients and their
medication: a post-discharge follow-up study. Age Ageing.
1992;21:178–181.
Conrad P. The meaning of medications: another look at compli-
ance. Soc Sci Med. 1985;20:29–37.
Craighead WE, Craighead LW, Ilardi SS. Psychosocial treatment
for major depressive disorder. In: Nathan PE, Garman SM,
eds. A Guide to Treatments That Work. New York: Oxford
University Press. 1998:226–239.
DiMatteo MR, Hays RD, Gritz ER, et al. Patient adherence to
cancer control regimens: scale development and initial vali-
dation. Psychological Assessment. 1993;5:102–112.
DiMatteo MR. Enhancing patient adherence to medical recom-
mendations. JAMA. 1994;271:79, 83.
DiMatteo MR. Variations in patients' adherence to medical rec-
ommendations: a quantitative review of 50 years of re-
search. Med Care. 2004 Mar;42:200–209.
Dunbar-Jacob J, Burke LE, Puczynski S. Clinical assessment and
management of adherence to medical regimens. In: Nicas-
sio PM, Smith TW, eds. Managing Chronic Illness: A Biopsy-
chological Perspective. CPP. Washington: American Psycho-
logical Association; 1995:313–349.
Eaddy M, Bramley T, Regan T. Time to antidepressant discontin-
uation: a comparison of controlled-release paroxetine and
immediate-release selective serotonin-reuptake inhibitors.
Manag Care Interface. 2003;16:22–27.
Finley PR, Rens HR, Pont JT, et al. Impact of a collaborative
pharmacy practice model on the treatment of depression in
primary care. Am J Health Syst Pharm. 2002;59:1518–1526.
Grahl C. Improving compliance: solving a $100 billion prob-
lems. Managed Healthcare. 1994;4:S11–S13.
Haynes RB, Sackett DL, Gibson ES, et al. Improvement of medi-
cation compliance in uncontrolled hypertension. Lancet.
1976;1:1265–1268.
CO M P L I A N C E
Katon W, Robinson P, Von Korff M, et al. A multifaceted inter-
vention to improve treatment of depression in primary
care. Arch Gen Psychiatry. 1996;53:1924–1932.
Katon W, Von Korff M, Lin E, et al. Collaborative management
to achieving treatment guidelines: impact upon depression
in primary care. JAMA. 1995;273:1026–1031.
Krupnick JL, Sotsky SM, Simmens S, et al. The role of the thera-
peutic alliance in psychotherapy and pharmacotherapy out-
come: findings in the National Institute of Mental Health
Treatment of Depression Collaborative Research Program. J
Consult Clin Psychol. 1996;64:532–539.
Leventhal H, Cameron L. Behavioral theories and the problem of
compliance. Patient Ed Counseling. 1987;10:117–138.
Lin EHB, Simon GE, Katon WJ, et al. Can enhanced acute-phase
treatment of depression improve long-term outcomes? A
report of randomized trials in primary care. Am J Psychia-
try. 1999;156:643–645.
Lin EHB, Von Korff M, Ludman EJ, et al. Enhancing adherence
to prevent depression relapse in primary care. Gen Hosp
Psychiatry. 2003;25:303–310.
Logan AG, Milne BJ, Achber C, et al. Work site treatment of
hypertension by specially trained nurses. A controlled trial.
Lancet. 1979;2:1175–1178.
Maguire P, Pitceathly C. Key communication skills and how to
acquire them. BMJ. 2002;325:697–700.
Rollman BL, Block MR, Schulberg HC. Symptoms of major de-
pression and tricyclic side effects in primary care patients. J
Gen Intern Med. 1997;12:284–291.
Sajatovic M, Ramirez LF. Rating Scales in Mental Health. Hud-
son, Ohio: Lexi-Comp. 2001.
Schulberg HC, Katon W, Simon GE, Rush AJ. Treating major de-
pression in primary care practice: an update of the Agency
for Health Care Policy and Research Practice Guidelines.
Arch Gen Psychiatry. 1998;55:1121–1127.
Schulberg HC, Raue PJ, Rollman BL. The effectiveness of
psychotherapy in treating depressive disorders in primary
care practice: clinical and cost perspectives. Review. Gen
Hosp Psychiatry. 2002;24:203–212.
Scott J. Using Health Belief Models to understand the efficacy-
effectiveness gap for mood stabilizer treatments. Neuropsy-
chobiology. 2002;46(suppl 1):13–15.
Unützer J, Katon W, Callahan CM, et al.Collaborative care man-
agement of late-life depression in the primary care setting:
a randomized controlled trial. JAMA. 2002;288:2836–2845.
Wells RA, Giannetti VJ. Handbook of the Brief Psychotherapies.
New York: Plenum Press. 1990.
DEPRESSION 47
 Q UA L I T Y M E A S U R E M E N T Q UA L I T Y M E A S U R E M E N T
A Review of HEDIS Measures and
Performance for Mental Disorders
BENJAMIN G. DRUSS, MD, MPH1
Emory University
SUMMARY
Performance on mental health HEDIS mea-
sures has been modest, with only minimal
improvements in recent years. Performance
on mental health measures has been consis-
tently worse than that for other medical
conditions. A critical step toward improving
performance is to understand where care is
provided and then to identify clinicians who
are responsible for ensuring that care is
delivered appropriately.
A
pproximately 35 million Americans experi-
ence a major depressive episode in their lives
(Kessler 2003). Depression is both highly
prevalent and extremely costly, placing a sig-
nificant economic burden on our health care system. In
2000, depression incurred an estimated $83.1 billion in
the United States in combined medical costs, lost pro-
ductivity, and premature mortality (Greenberg 2003).
There is a wide range of effective treatments for de-
pression, yet this disorder is underdiagnosed and under-
treated. Only about half of individuals with depression
are correctly diagnosed, and only about half of those
begun on treatment receive guideline-concordant care
(Hirschfeld 1997).
The majority are seen and treated in primary care set-
1Benjamin G. Druss, MD, MPH, is the Rosalynn Carter
Chair in Mental Health at the Rollins School of Public
Health of Emory University in Atlanta. Previously, he was
director of mental health policy studies at Yale University
Department of Psychiatry. He has been a member of the
NCQA HEDIS Behavioral Measures Development Sub-
group since 2002. Widely published, Druss has written nu-
merous papers on managed care and its implications for
mental health disorders. In recognition for his research, he
received the Academy Health Alice S. Hersh New Investi-
gator Award in 2003. He serves on the editorial boards of
General Hospital Psychiatry and Psychiatric Services.
48 P&T DIGEST
tings (Regier 1993). The
Agency for Health Care Pol-
icy and Research (AHCPR),
now the Agency for Health
Care Research and Quality
(AHRQ), developed guide-
lines to provide recommen-
dations for depression
treatment in primary care
settings (AHCPR 1993a,
AHCPR 1993b).
The AHCPR guidelines
serve as the basis for the de-
pression measures used in
the Health Plan Employer Benjamin G. Druss, MD, MPH
Data and Information Set,
the nation’s most widely used health care “report card.”
HEDIS rates health plans that cover 90 percent of Ameri-
cans enrolled in HMOs and is the set of quality indica-
tors used most by purchasers in selecting and rating
plans. Performance on this report card can provide im-
portant insights into plan quality as it is currently mea-
sured and reported in the U.S. health care marketplace.
HEDIS measures: evaluating performance
to improve outcomes
The National Committee for Quality Assurance
(NCQA) is an independent not-for-profit organization
dedicated to measuring and maintaining health care
quality through regular evaluation and accreditation
processes. HEDIS was developed in the early 1990s to
provide insight into the success of managed health care
plans in serving their constituents’ needs (NCQA 2000).
HEDIS data reflect performance in areas of health care
and service, including quality of care, access to care, and
member satisfaction with the plan and physicians.
NCQA selects indicators for HEDIS evaluation based
on their scientific soundness, relevance, feasibility, and
standardization as measures of health care provider per-
formance. Scientific soundness is ascertained based on
the accumulated literature supporting the accuracy, re-
producibility, and validity of the measure. The AHCPR
 Q UA L I T Y M E A S U R E M E N T

guidelines on depression in pri-

mary care provide the primary sci- TABLE 1 HEDIS: Five measures of mental health care quality
entific basis for the HEDIS de-
1. Percentage of members hospitalized for a mental disorder who had an am-
pression measures. Relevance is
bulatory visit with a mental health care provider within 7 days of discharge
determined by the measure’s po- 2. Percentage of members hospitalized for a mental disorder who had an am-
tential clinical and economic sig- bulatory visit with a mental health care provider within 30 days of discharge
nificance. Feasibility is defined by 3. Effective* treatment in the acute phase (ongoing in the 3-month period
the accessibility of the data and after a new depressive episode)
the nature of the costs for obtain- 4. Effective* continuation therapy (ongoing for 6 months after a new depres-
ing the data. Finally, standardiza- sive episode)
tion is the ability to create precise 5. Optimal practitioner contacts (at least three follow-up visits from a mental
specifications for measurement health care professional in the 3 months after a new depressive episode)
across multiple institutions. * Effective: NCQA defines this as ongoing treatment. SOURCE: DRUSS 2000
Five mental health care quality
measures are used in the HEDIS report card (Table 1). tion during the continuation phase of therapy. Follow-up
Two measures assess follow-up after discharge for any rates (based on the AHCPR recommended number of
mental disorder; three specifically assess depression contacts with a mental health care professional) decreased
medication management during the acute phase (first 3 from 21.4 percent in 1999 to 19.2 percent in 2002.
months), continuation phase (first 6 months), and ap- Mental health performance on HEDIS has consistently
propriate practitioner follow-up (at least three follow-up been worse than performance for other conditions. Com-
visits from a mental health care professional in the 3 pared with the HMOs’rate on nine nonmental health care
months after a new depressive episode). measures, mean rate of performance on mental health
care measures was much lower (48.0 vs 69.2 percent)
HEDIS PERFORMANCE (Druss 2002). Rates of improvement on mental health
Mental health performance on HEDIS measures based performance have also been less robust than for improve-
on data collected from 1999 to 2002 has been consistently ment on general medical domains (NCQA 2003).
modest (Figure 1), with only mini-
mal improvements during this pe-
riod (NCQA 2003). In commercial FIGURE 1 Increases in mental illness follow-up rates, 1999–2002
health plans, an average of 50 per- Percentage of patients in plan population, diagnosed with major depression, who
cent of patients with mental illness received follow-up care. Figures correspond with measures 1 and 2 in Table 1.
received follow-up care 7 days after
100
onset of a depressive episode. Fol-
low-up rates were even lower 90
among Medicaid and Medicare
populations,who had a less-than-40 80
percent rate of follow-up within the
70
week after discharge. As would be
expected, follow-up rates at 30 days 60
postdischarge were higher; nearly
75 percent of patients insured 50
through a commercial health plan, 40
and 60 percent insured through
Medicaid and Medicare, had a fol- 30
low-up visit during this period.
20
Table 2 shows antidepressant
treatment trends from 1999 to 2002
1999
2000
2001
2002

1999
2000
2001
2002

2000
2001
2002

2000
2001
2002

2000
2001
2002

2000
2001
2002

10
(NCQA 2003).About 60 percent of
patients received prescriptions dur- 0
7 days 30 days 7 days 30 days 7 days 30 days
ing the acute phase of an episode;
only about 40 percent were con- Commercial plans Medicaid Medicare
tacted by a mental health care pro- ADAPTED FROM NCQA 2003
fessional and renewed a prescrip-

DEPRESSION 49
 Q UA L I T Y M E A S U R E M E N T

IMPROVING QUALITY
TABLE 2 Management of antidepressants and number of visits OF DEPRESSION CARE
Percentage of eligible The health care system is pri-
patients in plan population* marily organized around provid-
HEDIS measure (corresponds with ing acute, episodic care, and is often
measures 3, 4, 5 in Table 1, page 49) 1999 2001† 2002 ill-suited for caring for persons with
Acute phase 58.8 56.9 59.8
chronic conditions. The most
Continuation phase 42.1 40.1 42.1 widely tested approach to improv-
Contacts 21.4 19.8 19.2 ing this care is the chronic care
†
model, developed by Edward Wag-
*Eligible: Plan members with diagnosed major depression. Data were not measured in 2000.
ADAPTED FROM NCQA 2003 ner at Group Health Cooperative of
Puget Sound (Wagner 1996). These
Why do plans score lower on mental health? interventions use multidisciplinary teams to follow pa-
A number of patient, provider, and system-level fac- tients with chronic diseases over time, actively engaging
tors are likely to explain the continuing low scores on them in managing their illnesses and ensuring that they fol-
HEDIS performance measures. In a recent study of pa- low up with treatment regimens. This model has been
tient and provider factors, Bull and colleagues surveyed shown to be effective in improving care for a range of
patients to understand the reasons for early discontinu- chronic illnesses, and it also has the potential to reduce costs
ation of SSRI medications. The study found miscom- of care (Bodenheimer 2002a, Bodenheimer 2002b).
munication between physicians and patients regarding More than a dozen studies have shown that the chronic
duration of therapy and side effects to be the most im- care approach can improve care for major depression in
portant cause of premature discontinuation (Bull 2002). primary care settings. In a recent review of the literature,
In a second study, Druss and colleagues examined plan- Gilbody (2002) found that whereas simple guidelines
level predictors of poor HEDIS mental health performance. and educational programs generally were inadequate,
Poor scores on general medical indices, failure to report complex programs that also included nurse case man-
findings publicly, and low medical-loss ratio (proportion agement and better integration between mental health
of revenues spent on clinical care) all predicted poor per- and primary care led to substantial improvements in de-
formance on the mental health measure (Table 3). pression care. Badamgarav and colleagues conducted a
A major challenge to understanding and improving meta-analysis of this literature (Figure 2) and found that
mental health performance among HMOs is the fact that such programs resulted in improved detection and treat-
many of these plans carve out their mental health care to ment of depression (Badamgarav 2003).
behavioral health managed care organizations (Frank
2003). Depression care may be provided solely within SUMMARY AND CONCLUSIONS
the HMO, within a carve-out, or in a combination of the While not perfect, HEDIS measures have placed qual-
two. A first critical step in plans improving their mental ity squarely on the table along with cost for health care
health performance is to understand where care is pro- purchasers and consumers. Scores on the mental health
vided, and then to identify the clinicians or clinics re- measures continue to lag those on other general health
sponsible for ensuring that it is delivered appropriately. indices, likely due to a complex combination of patient,

TABLE 3 HEDIS: Measures of HMOs’ mental health care performance*

Follow-up Appropriate use

Variable postdischarge of antidepressants
≥3 Acute Continuation
7 days 30 days contacts phase phase
HMOs performing given mental health care measure (%) 47.6 70.1 21.3 58.5 42.2
Correlates of poor mental health care performance (odds ratio)
Bottom quartile of HMOs on measures of quality medical care 3.15 4.03 2.56 2.04 1.56
Data not reported publicly 2.56 3.22 1.32 2.32 1.96
In bottom quartile of HMOs for medical-loss ratio 2.34 3.39 1.29 3.33 2.82
*N=384.
SOURCE: DRUSS 2002

50 P&T DIGEST
 Q UA L I T Y M E A S U R E M E N T

FIGURE 2 Features of disease management programs that influence

outcomes and processes of care

Detection of depression in a program with a screening component (N=1) . . . . . . . . . . . . . . . . . . . . . .

Outcomes influenced by providers’ and patients’ adherence (N=8). . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Patients’ satisfaction (N=6) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Adequacy of prescribed antidepressant treatment (N=11). . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Patients’ adherence to therapy (N=7) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Depressive symptoms (N=24) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Detection of depression in a program without a screening component (N=2) . . . . . . . . . . . . . . . . . . .
Referral to specialized care (N=2) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Health status (N=6) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Physical functioning (N=7) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Health care utilization (N=8). . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Hospitalization (N=2) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Health care costs (N=3) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
-1 -0.5 0 0.5 1
ADAPTED FROM BADAMGARAV 2003

provider, and system-level factors. A first step in im- Frank RG, Huskamp HA, Pincus HA. Aligning incentives in the
proving those scores is for plans to understand who is treatment of depression in primary care with evidence-
based practice. Psychiatric Services. 2003;54:682–687.
caring for patients with depression within their organi- Gilbody S, Whitty P, Grimshaw J, Thomas R. Educational and
zations, and to ensure that specific clinicians and clinics organizational interventions to improve the management
are held accountable for improving that care by follow- of depression in primary care: a systematic review. JAMA.
ing accepted guidelines for treatment. Organized 2003;289:3145–3151.
Greenberg PE, Kessler RC, Birnbaum HG, et al. The economic
approaches similar to those used for other chronic con- burden of depression in the United States: how did it
ditions are highly promising for improving the effec- change between 1990 and 2000? J Clin Psychiatry. 2003;
tiveness and cost effectiveness of depression care in the 64:1465–1475.
Hirschfeld RM, Keller MB, Panico S, et al. The National Depres-
United States. sive and Manic-Depressive Association consensus state-
ment on the undertreatment of depression. JAMA. 1997;
REFERENCES 277:333–340.
AHCPR (Agency for Health Care Policy and Research Depres- Kessler RC, Berglund P, Demler O, et al. The epidemiology of
sion Guideline Panel). Depression in Primary Care: Vol. 1. major depressive disorder: results from the National Co-
Detection and Diagnosis (Clinical Guideline Number 5). morbidity Survey Replication (NCS-R). JAMA. 2003;
Rockville, Md.: AHCPR. 1993a. 289:3095–3105.
AHCPR. Depression in Primary Care: Vol. 2. Treatment of Major NCQA (National Committee for Quality Assurance). HEDIS
Depression (Clinical Practice Guideline No. 5). Rockville, 2000: Technical Specifications. Vol 2. Washington: NCQA.
Md.: AHCPR. 1993b. 1999:105–110.
Badamgarav E, Weingarten SR, Henning JM, et al. Effectiveness Regier DA, Narrow WE, Rae DS, et al. The de facto US mental
of disease management programs in depression: a system- and addictive disorders service system. Epidemiologic
atic review. Am J Psychiatry. 2003;160:1080–1090. catchment area prospective 1-year prevalence rates of dis-
Bodenheimer T, Wagner EH, Grumbach K. Improving primary orders and services. Arch Gen Psychiatry. 1993;50:85–94.
care for patients with chronic illness. JAMA. 2002a;288: NCQA. The State of Health Care Quality 2003. Industry Trends
1775–1779. and Analysis. Washington: NCQA. 2003. Available at:
Bodenheimer T, Wagner EH, Grumbach K. Improving primary «http://www.ncqa.org/communications/
care for patients with chronic illness: the chronic care State%20Of%20Managed%20Care/
model, Part 2. JAMA. 2002b;288:1909–1914. SOHCREPORT2003.pdf». Accessed March 30, 2004.
Bull SA, Hu XH, Hunkeler EM, et al. Discontinuation of use and NCQA. The State of Managed Care Quality 2000. Washington:
switching of antidepressants: influence of patient-physician NCQA. 2000. Available at: «http://www.ncqa.org/
communication. JAMA. 2002;288:1403–1409. Communications/State%20Of%20Managed%20Care/
Druss BG, Miller CL, Rosenheck RA, et al. Mental health care SOMCReport2000.pdf». Accessed March 30, 2004.
quality under managed care in the United States: a view Wagner EH, Austin BT, Von Korff M. Organizing care for pa-
from the Health Employer Data and Information Set tients with chronic illness. Milbank Q. 1996;74:511–544.
(HEDIS). Am J Psychiatry. 2002;159:860–862.

DEPRESSION 51
 CO M O R B I D I T Y CO M O R B I D I T Y
Overview and Treatment
Of Social Anxiety Disorder
ASHOK RAJ, MD1
University of South Florida College of Medicine
SUMMARY
Social anxiety disorder is a highly prevalent
disorder that is frequently comorbid with
major depressive disorder. Pharmacothera-
pies, including SSRIs and benzodiazepines,
as well as cognitive-behavioral interven-
tions, are effective for many patients, helping
them to attain a higher quality of life.
S
ocial anxiety disorder (SAD) is the nation’s third
most prevalent psychiatric disorder, exceeded only
by substance abuse and major depressive disorder.
Though not recognized as a disorder until 1994
when the American Psychiatric Association (APA) pub-
lished the Diagnostic and Statistical Manual of Mental Dis-
orders, Fourth Edition (DSM-IV), SAD’s precursors were
well established in previous editions of DSM.
The U.S. lifetime prevalence of SAD among primary
care patients is 13.3 percent — 11.1 among men and 15.5
percent among women , with the mean age at onset of
SAD being 15.5 years (Magee 1996). With onset prior to
age 10, the condition tends to be severe. SAD is a chronic
condition with a mean duration of 19.4 years from the
time of diagnosis. It is somewhat more common among
women and — compared with unaffected persons, per-
sons with specific phobias or agoraphobia — SAD suf-
ferers are less likely to be married (Davidson 1993a).
CLINICAL PRESENTATION
Individuals with SAD have an excessive fear of being
scrutinized and negatively evaluated in social and per-
1 Ashok Raj, MD, is professor of psychiatry in the Depart-
ment of Psychiatry and Behavioral Medicine at the Uni-
versity of South Florida in Tampa. He is also director of the
Division of Geriatric Psychiatry at USF College of Medi-
cine. He is board certified in general psychiatry with added
qualifications in geriatric psychiatry. His research interests
include anxiety and depressive disorders. He earned his
MBBS from Madras Medical College in India.
52 P&T DIGEST
formance situations. They
also recognize that their
fears are excessive, interfer-
ing, and distressing. Those
with SAD may experience
significant anticipatory
anxiety before entering the
feared situation and suffer
panic attacks when in the
phobic situation. They may
cope through avoidance.
SAD is classified as non-
generalized (NGSAD) when
fewer than three phobias are
present; fear of public speak- Ashok Raj, MD
ing is the most frequent non-
generalized phobia. In two thirds of all SAD cases, the
person has many of the following multiple avoidances: in-
teracting with strangers or persons in authority; being as-
sertive and refusing unreasonable requests; confronting the
behaviors of others; dating; attending parties; using pub-
lic restrooms; eating or working while being watched.
Persons with the generalized subtype of SAD (GSAD)
tend to be more functionally disabled (Kessler 1998),
experience onset at an earlier age, be less educated, and
unemployed (Heimberg 1990b). Children who consis-
tently respond to novel or unfamiliar situations with be-
havioral inhibition may be predisposed to developing
SAD later in life (Kagan 1988). The prevalence of SAD
in parents of children predisposed to developing the dis-
order is 18 percent; in contrast, only 3 percent of parents
of noninhibited children are affected (Rosenbaum 1994).
Differential diagnosis and diagnostic tools
SAD differential diagnoses include the other anxiety
disorders, particularly panic disorder. The diagnostician
should also consider other conditions, such as avoidant
personality disorder, shyness, or psychosis.
Two reliable, easy-to-administer tools are available to
help to diagnose SAD:
• Module G of the Mini International Neuropsychi-
atric Interview (MINI) (Figure 1) uses five screen-

ing questions to identify a diagnosis of SAD (Shee-
han 1997, Sheehan 1998).
• The Liebowitz Social Anxiety Scale (LSAS)2 is a
clinician-rated scale that covers 24 feared social and
performance situations. Each phobic situation listed
is rated for the intensity of fear it provokes and the
frequency with which the individual avoids it. LSAS
offers a convenient way to monitor patient progress
during treatment (Liebowitz 1987, Heimberg 1999).
Consequences and cormorbidity
Comorbidity with SAD is 69 percent (most commonly:
specific phobia, 59 percent; agoraphobia, 45 percent)
(Schneier 1992b). With comorbid psychiatric conditions,
the prognosis for recovery decreases (Davidson 1993a).
Comorbidity increases the likelihood of a suicide attempt
15-fold. Rates of alcohol abuse range from 15 to 27 percent
(Schneier 1992b).
GSAD tends to be more disabling than NGSAD; two
thirds of sufferers are single or divorced, half have com-
pleted high school; and 1 in 5 is on disability (Kessler
1998). These patients utilize more outpatient medical
treatment than do controls. Compared with those with
the pure disorder, those with comorbid psychiatric con-
ditions are more likely to seek treatment (Schneier 1992b).
TREATMENT
Treatment options include pharmacotherapy and be-
havioral interventions. Selective serotonin reuptake in-
hibitors (SSRIs) and high-potency benzodiazepines are
medications of choice. The first step toward developing
an effective treatment plan is to make the diagnostic dis-
tinction between GSAD and NGSAD, as therapeutic op-
tions tend to differ. Given that SAD has an early age of
onset, treatment should begin ideally during adoles-
cence, as early intervention may halt progression and
prevent such complications as depression, panic disor-
der, substance abuse, and suicide.
Pharmacologic options
Pharmacotherapy is an important first step in treating
most GSAD patients. Evidence shows considerable effi-
cacy for some SSRIs, reversible inhibitors of monoamine
oxidase (RIMAs), monoamine oxidase (MAO) in-
hibitors, and some benzodiazepines (Table 1, page 54).
These classes of medication differ, and one way to
compare their efficacy is to assess their weighted effect
size (ES). This kind of analysis detects differences be-
tween the proportions of response to the drug versus
2 The LSAS is widely available online. A printable version is ac-
cessible at «http://healthnet.umassmed.edu/mhealth/
LiebowitzSocialAnxietyScale.pdf»; an interactive version is ac-
cessible at «http://www.psychmeds.com/liebowitz.html».
CO M O R B I D I T Y
placebo, adjusted for sample size. ES ratings for selected
classes of SAD therapeutics, in descending order, are
benzodiazepines, 1.00 (van Vliet 1994, Davidson 1993b);
phenelzine, 0.98 (Hidalgo 2001); anticonvulsants, 0.46
(Davidson 2003); RIMAs, 0.41 (Davidson 2003); bus-
pirone and beta blockers, 0.14 (Davidson 2003).
To date, paroxetine and sertraline are the only two
medications approved by the U.S. Food and Drug Ad-
ministration (FDA) for the treatment of GSAD.
SSRI therapy
Paroxetine. In an 11-week, open-label study, 36 GSAD
patients were treated with paroxetine and rated, at com-
pletion, using the Clinical Global Improvement (CGI)
scale. Of 30 patients who completed the study, 23 were
much improved or very much improved. The mean ef-
fective dosage of paroxetine was 47.9 mg/day (range:
20–50 mg). Sixteen subjects who continued to participate
in a double-blind, relapse-prevention study were ran-
domly assigned either to continue receiving paroxetine
or to receive placebo for another 12 weeks. Of the eight
patients in the placebo group, five relapsed; 1 of 8 in the
paroxetine group relapsed (Stein 1996). Efficacy of parox-
etine also was established in a 12-week, multicenter,
double-blind, placebo-controlled trial in which 187 pa-
tients were randomly assigned to paroxetine or placebo.
At the clinician’s discretion, the dosage was titrated by in-
creasing it by 10 mg/week to a maximum of 50 mg/day.
At a mean effective dosage of 36.6 mg/day, 55 percent of
the paroxetine group and 24 percent of the placebo group
were either much improved or very much improved on
the CGI scale. The paroxetine group also showed signif-
icantly greater improvement on the LSAS (Figure 2) and
on the social disability and work disability subscales of
the Sheehan Disability Scale (Stein 1998). Further sup-
port for paroxetine in treating GSAD patients derives
from a study in which subjects randomized to paroxetine
had a 70 percent response rate compared with an 8 per-
cent response rate with placebo (Allgulander 1999).
Sertraline. In a double-blind, placebo-controlled,
crossover study of sertraline in GSAD, 12 patients re-
ceived either sertraline or placebo for 10 weeks, no treat-
ment for 2 weeks, and subsequently the other treatment
for an additional 10 weeks. Starting at 50 mg/day, the
dosage was increased by 50 mg every 2 weeks, at the clin-
ician’s discretion. The researchers established the mean
effective dosage of sertraline to be 133.5 mg/day. They
found a statistically significant reduction in the mean
LSAS score in 50 percent of patients receiving sertraline
and only 9 percent receiving placebo (Katzelnick 1995).
In a larger, more recent, double-blind, placebo-controlled
study, investigators found a 53 percent response rate (71
of 135) for sertraline and a 29 percent response rate (20
DEPRESSION 53
 CO M O R B I D I T Y

cians will need to extrapolate from

TABLE 1 Controlled studies of selected pharmacotherapies studies in adult populations.
for social anxiety disorder Other SSRIs. Even though flu-
Placebo voxamine has not been approved
Response response Mean by the FDA for the treatment of
Drug rate (%) rate (%) Reference daily dose SAD, there is credible evidence for
Antiepileptic
its efficacy. A randomized 12-week,
Gabapentin 38 17 Pande 1999 2,100 mg double-blind, placebo-controlled
Anxiolytic study compared fluvoxamine 50–
Buspirone 7 7 van Vliet 1997 30.0 mg 150 mg/day, with placebo in the
Benzodiazepines treatment of GSAD. Of 30 patients
Alprazolam 38 20 Gelertner 1991 4.2 mg in the study, 46 percent responded
Bromazepam 82 20 Versiani 1997 21.0 mg to fluvoxamine and 7 percent re-
Clonazepam 78 20 Davidson 1993 2.4 mg sponded to placebo, with reduced
Beta blocker LSAS scores. In the 12-week con-
Atenolol 30 23 Liebowitz 1992 97.6 mg tinuation phase, 14 of 16 fluvox-
MAO inhibitors
amine patients improved further
Phenelzine 69 20 Gelertner 1991 55.0 mg
Phenelzine 64 23 Liebowitz 1992 75.7 mg
(van Vliet 1994). Another double-
Phenelzine 65 33 Heimberg 1998 60.0 mg blind, placebo-controlled multi-
Phenelzine 85 15 Versiani 1992 67.5 mg center study replicated findings on
RIMAs fluvoxamine’s efficacy and safety
Brofaromine 50 19 Lott 1997 107 mg in the acute treatment of GSAD,
Brofaromine 78 23 Fahlen 1995 150 mg with a final mean effective dosage
Brofaromine 26 0 van Vliet 1992 150 mg of 202 mg/day (Stein 1999).
Moclobemide 65 15 Versiani 1992 570.7 mg In dosages ranging 20–60 mg/
Moclobemide 38 33 Noyes 1997 900 mg day, fluoxetine has shown efficacy in
Moclobemide 17 13 Schneier 1998 728 mg four case series reports of 2 patients
Moclobemide 47 34 IMCTG 1997 600 mg, P<.05;
(Sternbach 1990), 12 patients (Sch-
300 mg, NS
SSRIs
neier 1992a), 14 patients (Black
Fluvoxamine 46 7 van Vliet 1994 50–150 mg 1992), and 16 patients (Van Amer-
Fluvoxamine 43 23 Stein 1999 202 mg ingen 1993). An open trial of venla-
Paroxetine 55 24 Stein 1998 36.6 mg faxine in 10 patients shows promis-
Paroxetine 70 8 Allgulander 1999 35.0 mg ing results (Emmanuel 1995).
Paroxetine 66 32 Baldwin 1999 35.0 mg Extended-release venlafaxine is
Sertraline 50 9 Katzelnick 1995 133.5 mg under study, and citalopram has not
Sertraline 53 29 Van Ameringen 2001 125 mg yet been tested for relief of SAD.
Tricyclic antidepressant While no SSRI study has included
Imipramine * * Zitrin 1983 * either another SSRI or non-SSRI di-
MAO=monoamine oxidase inhibitor; NS=not significant; RIMA=reversible inhibitors of monoamine rect, active comparator, expert con-
oxidase; SSRI=selective serotonin reuptake inhibitor. sensus favors SSRIs to alternatives,
*In a controlled-outcome study involving 218 adult phobic patients, imipramine was superior to
placebo in patients with spontaneous panic attacks (patients with agoraphobia or mixed pho- as they are safer than the MAO in-
bia). In patients with simple phobia, who do not experience spontaneous panic, there was no sig- hibitors and less likely to cause de-
nificant difference between imipramine and placebo. pendence than the benzodiazepines.

of 69) for placebo; sertraline responders averaged an im- Benzodiazepine therapy

provement of 17 points (reduction from baseline) on the
Brief Social Phobia Scale, while placebo responders’ av-
erage improvement was 9 points (Van Ameringen 2001).
An open-label study showed sertraline, at mean
dosages of 123 mg per day, to be effective in children age
10 to 17 (Compton 2001). Several large international
clinical trials of SAD therapy among children and ado-
lescents are continuing. Until results are available, clini-
To date, two double-blind, placebo-controlled studies
and nine open-label studies of benzodiazepines indicate
efficacy for this drug class. Open-label studies suggest
that clonazepam (Ontiveros 1990), alprazolam (Reich
1988), and bromazepam (Versiani 1997) are effective in
many cases. The double-blind studies suggest that clona-
zepam (Davidson 1993b) but not alprazolam (Gelertner
1991) is superior to placebo. In a 10-week study of 75 pa-

54 P&T DIGEST
 CO M O R B I D I T Y

tients, 78 percent of the clonaze-

pam group but only 20 percent of
the placebo group responded to
therapy. The mean dosage of clon-
azepam at endpoint was 2.4 mg/
day (Davidson 1993b).
Other classes
Drugs from other classes, in-
cluding MAO inhibitors, RIMAs,
tricyclic antidepressants, and beta
blockers have been studied in the
treatment of SAD. Table 1 lists re-
sponse rates from selected placebo-
controlled studies.
FIGURE 1 Module G, Mini International Neuropsychiatric Interview
G. SOCIAL PHOBIA (Social Anxiety Disorder)
( ➔ means: go to the diagnostic box, circle NO and move to the next module)
G1. In the past month, were you fearful or embarrassed being watched, being
the focus of attention, or fearful of being humiliated? This includes things
like speaking in public, eating in public or with others, writing while
someone watches, or being in social situations.
No ➔ Yes
G2. Is this fear excessive or unreasonable?
No ➔ Yes
G3. Do you fear these situations so much that you avoid
them or suffer through them? No Yes

No ➔ Yes SOCIAL PHOBIA

Treatment response
Response is associated with du- G4. Does this fear disrupt your normal work or social
ration of therapy, which has been functioning or cause you significant distress?
minimally studied, resulting in a
dearth of useful data. In one study,
patients with SAD who responded SOURCE: SHEEHAN 1997, SHEEHAN 1998
to therapy continued to improve
beyond the 8th week of active treatment (van Vliet 1994).
Poor response is associated with symptom severity at
baseline, family history of SAD, higher systolic blood
pressure, abnormal heart rate, and alcohol consumption.
Relapse is high following discontinuation of successful
pharmacotherapy. Patients suffered a relapse rate of 20
percent when they switched from clonazepam to placebo
(Gelertner 1991); 62 percent of those switching from
paroxetine to placebo relapsed, compared with 12 per-
cent of patients continuing paroxetine therapy (Stein
1996) or switching to sertraline (Van Ameringen 2001).
Nonpharmacologic options
Behavior therapy often helps patients overcome social
anxiety. Exposure therapy has been shown to be effective
(Fava 1989), and the treatment of choice for NGSAD is
in vivo exposure behavior therapy. It eliminates excessive
fear by placing the patient, repeatedly and for prolonged
periods, in the specific situation that evokes the phobic
response. Self-help groups — such as Toastmasters In-
ternational, a global network of clubs that foster the de-
velopment of public speaking in a practicum environ-
ment — are useful and recommended.
Adjunctively, in some cases, a beta blocker — for ex-
ample, propranolol, 10 to 40 mg, taken 1 to 2 hours be-
fore a specific nongeneralized performance anxiety sit-
uation, may reduce tachycardia or hand tremor, though
there have been no controlled trials to support it. Fur-
ther, studies investigating beta blockers in GSAD have
found that they are not superior to placebo (Liebowitz
(Social Anxiety
Disorder)
CURRENT
No ➔ Yes
1992). In a study by Turner (1994), flooding, a form of
behavior therapy, was found to be more effective than
atenolol (62 vs. 38 percent).
Other behavior-therapy studies have found exposure
therapy to be superior to placebo (Turner 1994), waiting-
list controls (Butler 1984, Newman 1994), and progres-
sive relaxation therapy (Al-Kubaisy 1992, Alström 1984).
Most evidence suggests that therapist-guided exposure
is superior to exposure without guidance, in which case
patients cope alone with their social anxiety (Mattick
1989). Phobic patients tend to be noncompliant with ex-
posure instructions, and compliance with exposure in-
structions between sessions has been linked with better
outcomes immediately following treatment (Leung 1996)
and at 6-month follow-up (Edelman 1995).
Combination treatment — cognitive restructuring
plus behavior therapy — was found to be superior in pa-
tients receiving both educational/supportive therapy (He-
imberg 1990a) and in waiting-list control groups
(DiGiuseppe 1990, Hope 1995). Yet the debate continues
over whether cognitive techniques enhance efficacy com-
pared with exposure alone. Some studies show equivalent
results for cognitive restructuring plus exposure, and for
exposure alone (Feske 1995, Gould 1997), while others
find only cognitive restructuring combined with exposure
to be significantly superior to placebo (Taylor 1996). In
a recent meta-analysis, researchers compared five CBTs —
exposure therapy, cognitive restructuring, exposure with
cognitive restructuring, social skills training, and applied
relaxation — and found all to be moderately effective for

DEPRESSION 55
 CO M O R B I D I T Y
FIGURE 2 Effectiveness of paroxetine in social anxiety
disorder compared with placebo*
0 Placebo
–5
Improvement in LSAS score
–10
–15
–20
–25
–30
–35
–40
0 2
LSAS=Liebowitz Social Anxiety Scale.
*P<.05 vs placebo for paroxetine at all values, weeks 2 through 12.
SOURCE: STEIN 1998
SAD, with no significant differences between them im-
mediately following or at later follow-up (Federoff 2001).
Individual and treatment-group formats appear to be
equally effective. The results of social-skills training to
treat SAD are mixed (Falloon 1981, Lucock 1988). The
only controlled study of social-skills training, applied
over 15 weeks, found no significant difference between
treated patients and a control group (Marzillier 1976).
Progressive muscle-relaxation training has shown little
efficacy in the treatment of SAD (Alström 1984), and
there are no controlled studies of psychotherapy in SAD.
SUMMARY
In the past decade, we have learned that GSAD is much
more common and disabling than previously realized.
This chronic condition has significant comorbidity and
increased risk for substance abuse/dependence and sui-
cide. Current data indicate that several pharmacologic
and behavioral treatment options are effective in easing
symptoms, alleviating disabilities, and mitigating the
complications from which these patients suffer.
REFERENCES
Al-Kubaisy T, Marks IM, Logsdail S, et al. Role of exposure
homework in phobia reduction: a controlled study. Behav
Ther. 1992;23:599–621.
Allgulander C. Paroxetine in social anxiety disorder: a random-
ized placebo-controlled study. Acta Psychiatr Scand.
1999;100:193–198.
56 P&T DIGEST
Alström JE, Nordlund CL, Persson G, et al. Effects
of four treatment methods on social phobic
patients not suitable for insight-oriented
psychotherapy. Acta Psychiatr Scand.
1984;70:97–110.
Baldwin D, Bobes J, Stein DJ, et al. Paroxetine in
social phobia/social anxiety disorder. Ran-
Paroxetine domized double-blind, placebo-controlled
study. Paroxetine Study Group. Br J Psychia-
try. 1999;175:120–126.
Beaumont G. A large open multicentre trial of
clomipramine (Anafranil) in the manage-
ment of phobic disorders. J Int Med Res.
1977;5(suppl 5):116–123.
Black B, Uhde TW, Tancer ME. Fluoxetine for
treatment of social phobia (letter). J Clin Psy-
chopharmacol. 1992;12:293–295.
Butler G, Cullington A, Munby M, et al. Exposure
and anxiety management in the treatment of
social phobia. J Consult Clin Psychol.
1984;52:642–650.
Compton SN, Grant PJ, Chrisman AK, et al. Ser-
traline in children and adolescents with so-
cial anxiety disorder: an open trial. J Am Acad
4 6 8 10 12
Child Adolesc Psychiatry. 2001;40:564–571.
Week Davidson JR. Pharmacotherapy of social phobia.
Acta Psychiatr Scand Suppl. 2003;417:65–71.
Davidson JR, Hughes DL, George LK, Blazer DG.
The epidemiology of social phobia: findings
from the Duke Epidemiological Catchment
Area Study. Psychol Med. 1993a;23:709–718.
Davidson JR, Potts NL, Richichi E, et al. Treatment of social pho-
bia with clonazepam and placebo. J Clin Psychopharmacol.
1993b;13:423–428.
DiGiuseppe R, McGowan L, Sutton-Simon K, Gardner F. A com-
parative outcome study of four cognitive therapies in the
treatment of social anxiety. J Rational-Emotive Cog-Behav
Ther. 1990;8:129–146.
Edelman RE, Chambless DL. Adherence during sessions and
homework in cognitive-behavioral group treatment of so-
cial phobia. Behav Res Ther. 1995;33:573–577.
Emmanuel NP, Czepowicz VD, Villareal G, et al. Venlafaxine in
social phobia: a case series. New research poster #113, pre-
sented at the 148th Annual Meeting of the American Psy-
chiatric Association, Miami, May 20–25, 1995.
Fahlen T, Nilsson HL, Borg K, et al. Social phobia: the clinical ef-
ficacy and tolerability of the monoamine oxidase-A and
serotonin uptake inhibitor brofaromine. A double-blind
placebo-controlled study. Acta Psychiatr Scand. 1995;92:
351–358.
Falloon IR, Lloyd GG, Harpin RE. The treatment of social pho-
bia. Real-life rehearsal with nonprofessional therapists. J
Nerv Ment Dis. 1981;169:180–184.
Fava GA, Grandi S, Canestrari R. Treatment of social phobia by
homework exposure. Psychother Psychosom. 1989;52:
209–213.
Federoff IC, Taylor S. Psychological and pharmacological treat-
ments of social phobia: a meta analysis. J Clin Psychiatry.
2001;21:311–324.
Feske U, Chambless DL. Cognitive behavioral versus exposure
only treatment for social phobia: a meta-analysis. Behav
Ther. 1995;26:695–720.
Gelertner CS, Uhde TW, Cimbolic P, et al. Cognitive-behavioral
and pharmacological treatments of social phobia. Arch Gen
Psychiatry. 1991;38:938–945.
Gould RA, Buckminster S, Pollack MH, et al. Cognitive-behav-
ioural and pharmacological treatment for social phobia: a

meta-analysis. Clin Psychol Sci Pract. 1997;4:291–306.
Heimberg RG, Dodge CS, Hope DA, et al. Cognitive-behavioral
group treatment of social phobia: comparison to a credible
placebo control. Cogn Ther Res. 1990a;14:1–23.
Heimberg RG, Homer KJ, Juster HR, et al. Psychometric proper-
ties of the Liebowitz Social Anxiety Scale. Psychol Med.
1999;29:199–212.
Heimberg RG, Hope DA, Dodge CS, et al. DSM-III-R subtypes
of social phobia. Comparison of generalized social phobics
and public speaking phobics. J Nerv Ment Dis. 1990b;178:
172–179.
Heimberg RG, Liebowitz MR, Hope DA, et al. Cognitive behav-
ioral group therapy vs phenelzine therapy for social pho-
bia: 12-week outcome. Arch Gen Psychiatry. 1998;55:
1133–1141.
Hidalgo RB, Barnett SD, Davidson JR. Social anxiety disorder:
two decades of progress. Int J Neuropsychopharmacol.
2001;4:279–298.
Hope DA, Heimberg RG, Bruch MA. Dismantling cognitive-
behavioral group therapy for social phobia. Behav Res Ther.
1995;33:637–650.
IMCTG (International Multicenter Clinical Trial Group) on
moclobemide in social phobia. A double blind, placebo
controlled clinical study. Eur Arch Psychiatry Clin Neurosci.
1997;247:71–80.
Kagan J, Resnick JS, Snidman N. Biological basis of childhood
shyness. Science. 1988;240:167–171.
Katzelnick DJ, Kobak KA, Greist JH, et al. Sertraline for social
phobia: a double-blind, placebo-controlled crossover study.
Am J Psychiatry. 1995;152:1368–1371.
Kessler RC, Stein MB, Berglund P. Social phobia subtypes in the
National Comorbidity Survey. Am J Psychiatry. 1998;155:
613–619.
Leung AW, Heimberg RG. Homework compliance, perceptions
of control and outcome of cognitive-behavioral treatment
of social phobia. Behav Res Ther. 1996;34:423–432.
Liebowitz MR, Schneier FR, Campeas R, et al. Phenelzine vs
atenolol in social phobia. A placebo-controlled compari-
son. Arch Gen Psychiatry. 1992;49:290–300.
Liebowitz MR. Social phobia. Mod Probl Pharmacopsychiatry.
1987;22:141–173.
Lott M, Greist JH, Jefferson JW, et al. Brofaromine for social
phobia: a multicenter, placebo-controlled, double-blind
study. J Clin Psychopharmacol. 1997;17:255–260.
Lucock MP, Salkovskis PM. Cognitive factors in social anxiety
and its treatment. Behav Res Ther. 1988;26:297–302.
Magee WJ, Eaton WW, Witchen HU, et al. Agoraphobia, simple
phobia, and social phobia in the National Comorbidity
Survey. Arch Gen Psychiatry. 1996;53:159–168.
Marzillier JS, Lambert C, Kellet J. A controlled evaluation of sys-
tematic desensitization and social skills training for socially
inadequate psychiatric patients. Behav Res Ther. 1976;14:
225–238.
Mattick RP, Peters L, Clarke JC: Exposure and cognitive restruc-
turing for social phobia: a controlled study. Behav Ther.
1989;20:3–23.
Newman MG, Hofmann SG, Trabert W, et al. Does behavioral
treatment of social phobia lead to cognitive changes? Behav
Ther. 1994;25:503–517.
Noyes R Jr, Moroz G, Davidson JT, et al. Moclobemide in social
phobia: a controlled dose-response trial. J Clin Psychophar-
macol. 1997;17:247–254.
Ontiveros A, Fontaine R. Social phobia and clonazepam. Can J
Psychiatry. 1990;35:439–441.
Pande AC, Davidson JR, Smith RD, Jefferson JW. Treatment of
social phobia with gabapentin: a placebo-controlled study.
J Clin Psychopharmacol. 1999;19:341–348.
CO M O R B I D I T Y
Reich J, Yates W. A pilot study of treatment of social phobia with
alprazolam. Am J Psychiatry. 1988;145:590–594.
Rosenbaum JF, Beiderman J, Pollock RA, Hirshfeld DR. The eti-
ology of social phobia. J Clin Psychiatry. 1994;55(suppl):
10–16.
Schneier FR, Chin SJ, Hollander E, Liebowitz MR. Fluoxetine in
social phobia (letter). J Clin Psychopharmacol.
1992a;12:62–64.
Schneier FR, Goetz D, Campeas R, et al. Placebo-controlled trial
of moclobemide in social phobia. Br J Psychiatry.
1998;172:70–77.
Schneier FR, Johnson J, Hornig CD, et al. Social phobia: comor-
bidity and morbidity in an epidemiologic sample. Arch Gen
Psychiatry. 1992b;49:282–288.
Sheehan DV, Lecrubier Y, Hartnett Sheehan K, et al. Reliability
and validity of the MINI International Neuropsychiatric
Interview (MINI): according to the SCID-P. Eur Psychiatry.
1997;12:232–241.
Sheehan DV, Lecrubier Y, Sheehan KH, et al. The Mini-
International Neuropsychiatric Interview (MINI): the de-
velopment and validation of a structured diagnostic psychi-
atric interview for DSM-IV and ICD-10. J Clin Psychiatry.
1998;59(suppl 20):22–33.
Stein MB, Chartier MJ, Hazen AL, et al. Paroxetine in the treat-
ment of generalized social phobia: open-label treatment
and double-blind, placebo-controlled discontinuation. J
Clin Psychopharmacol. 1996;16:218–222.
Stein MB, Fyer AJ, Davidson JR, et al. Fluvoxamine treatment of
social phobia (social anxiety disorder): a double-blind,
placebo-controlled study. Am J Psychiatry. 1999;156:
756–760.
Stein MB, Liebowitz MR, Lydiard RB, et al. Paroxetine treatment
of generalized social phobia (social anxiety disorder): a ran-
domized controlled trial. JAMA. 1998;280:708–713.
Sternbach H. Fluoxetine treatment of social phobia. Clin Psy-
chopharmacol. 1990;10:230–231.
Taylor S. Meta-analysis of cognitive-behavioral treatments for
social phobia. J Behav Ther Exp Psychiatry. 1996;27:1–9.
Turner SM, Beidel DC, Jacob RG. Social phobia: a comparison of
behavior therapy and atenolol. J Consult Clin Psychol.
1994;62:350–358.
Van Ameringen MA, Lane RM, Walker JR, et al. Sertraline treat-
ment of generalized social phobia: a 20-week, double-blind,
placebo-controlled study. Am J Psychiatry.
2001;158:275–281.
Van Ameringen M, Mancini C, Streiner DL. Fluoxetine efficacy
in social phobia. J Clin Psychiatry. 1993;54:27–32.
van Vliet IM, den Boer JA, Westenberg HG. Psychopharmacolog-
ical treatment of social phobia: clinical and biochemical ef-
fects of brofaromine, a selective MAO-A inhibitor. Eur
Neuropsychopharmacol. 1992;2:21–29.
van Vliet IM, den Boer JA, Westenberg HG. Psychopharmacolog-
ical treatment of social phobia: a double-blind placebo-
controlled study with fluvoxamine. Psychopharmacology
(Berl). 1994;115:128–134.
van Vliet IM, den Boer JA, Westenberg HG, Pian KL. Clinical ef-
fects of buspirone in social phobia: a double-blind placebo-
controlled study. J Clin Psychiatry. 1997;58:164–168.
Versiani M, Nardi AE, Figueria I, et al. Double blind placebo
controlled trial with bromazepam in social phobia. J Bras
Psiq. 1997;46:167–171.
Versiani M, Nardi AE, Mundim FD, et al. Pharmacotherapy of
social phobia. A controlled study with moclobemide and
phenelzine. Br J Psychiatry. 1992;161:353–360.
Zitrin CM, Klein DF, Woerner MG, et al. Treatment of phobias.
I. Comparison of imipramine hydrochloride and placebo.
Arch Gen Psychiatry. 1983;40:125–138.
DEPRESSION 57
 CONTINUING EDUCATION POST-TEST
P&T DIGEST Depression
On the combined answer sheet/evaluation form on page 60, please place an X through the box of the letter
corresponding with the correct response for each question. There is only one correct answer to each question.

1. In the United States, the lifetime risk of major 6. Lifetime prevalence of social anxiety disorder
depression is: (SAD) in the United States is:
a. 1–3 percent. a. 7.3 percent.
b. 10.3 percent. b. 11.1 percent.
c. 15 percent. c. 13.3 percent.
d. 17.3 percent. d. 15.5 percent.
2. Which is true of practice guidelines for depression 7. The mechanism of action of SSRIs and SNRIs may
in primary care? be associated with alterations in gene expression
a. They are known to be widely used. that produce sustained changes in the function of
b. Pharmacoeconomic studies show they improve selected brain cells.
outcomes. a. True.
c. Use and careful interpretation of guidelines im- b. False.
proves detection and reduces relapse and suicide
8. During the past 2 decades, pharmacoeconomic
risk.
literature regarding depression management has
d. None are available.
generally focused on all but which of the
3. For selective serotonin reuptake inhibitors (SSRIs) following questions:
or serotonin norepinephrine reuptake inhibitors a. Are neurotransmitter receptor-specific anti-
(SNRIs) to produce an antidepressant response: depressants cost-effective alternatives to older
a. Increasing or reducing brain levels of serotonin agents?
only is probably sufficient. b. Are there differences in clinical and economic out-
b. Increasing or reducing brain levels of serotonin comes between the newer antidepressants?
only is probably insufficient. c. What are the direct and indirect costs of patients’
c. Increasing and reducing levels of serotonergic, nor- noncompliance with antidepressant therapy?
adrenergic, and other neurotransmitter activity d. Does compliance with clinical guidelines regarding
may be needed. the recommended length of antidepressant ther-
d. Answers a and c only. apy translate into beneficial economic outcomes?
e. Answers b and c only.
9. In primary care, which costs more?
4. Poor compliance with antidepressant therapy can a. Diagnosis and treatment of depression.
result from: b. Subsequent laboratory costs searching for other di-
a. Miscommunication between practitioner and pa- agnoses.
tient.
10. An AHRQ national initiative found that __ percent
b. Lack of collaboration between the practitioner and
of patients with depression received appropriate
patient.
care in primary care settings.
c. Patient-perceived stigma of depression and anti-
a. 25.
depressants.
b. 30.
d. All the above.
c. 35.
5. To evaluate follow-up care after a new depressive d. 40.
episode, the Health Plan Employer Data and
11. For dysthymia and acute major depression, the
Information Set (HEDIS) report card consists of
American College of Physicians–American Society
several key measures, including:
of Internal Medicine recommends drug treatment
a. Effective therapy in initial 3-month period.
at the same dose for at least:
b. Effective continuation therapy for 6 months.
a. 4 weeks.
c. At least 3 follow-up visits from a mental health care
b. 12 weeks.
professional in 3 months.
c. 4 months.
d. Answers a and c only.
d. 6 months.
e. Answers a through c.

58 P&T DIGEST
 CONTINUING EDUC ATION

12. Components of effective adherence counseling 19. Total health care costs for patients older than 65
are: with significant symptoms of depression are
a. Empathy with patients’ wishes and individualism. approximately ___ percent more than for non-
b. Explaining how treatment benefits outweigh barri- depressed elderly patients.
ers. a. 30.
c. Addressing patients’ concerns about side effects. b. 40.
d. Explaining the neurobiologic basis of depression. c. 50.
e. All the above. d. 60.
13. The following characteristics are typical of 20. All are true of professional society guidelines
individuals with generalized SAD: except:
a. Early-age onset. a. There is a 3–5 year gap between updates.
b. Less education. b. Preparation involves many steps and checks.
c. Unemployment. c. They must be peer reviewed.
d. Answers a and b only. d. They are mandates that practitioners must follow in
e. Answers a through c. order to be reimbursed.
14. Using a human-capital approach, the annual cost 21. Increasing length of antidepressant therapy
of depression in the United States has been to a duration consistent with national clinical
estimated at $____ billion. guidelines translates into a reduction in total
a. 29.7. health care expenditures.
b. 34.8. a. True.
c. 43.7. b. False.
d. 51.2.
22. After 6 months, what effect did controlled-release
15. Effective practice guidelines: formulations of SSRIs have on compliance with
a. Define the clinical setting for which the guidelines antidepressant therapy?
are appropriate. a. Significantly increased compliance.
b. Consider all important treatment options and their b. Decreased compliance.
consequences.
23. Regarding health care performance as measured
c. Clearly recommend interventions.
by HEDIS, which of the following statements is
d. Devise a mechanism for dissemination, evaluation,
false?
and updates.
a. HMO performance is substantially lower on mental
e. All of the above.
health care measures versus nonmental health care
f. Answers b, c, and d only.
measures.
16. Which is an inherent limitation of randomized b. High medical-loss ratio also is associated with poor
clinical trials with an economic component? performance.
a. Inability to control confounders. c. HMOs that do not report their quality of care data
b. Inability to generalize to a broader population. are more likely to perform poorly on the HEDIS
c. Data are subject to publication and/or selection mental health care assessment.
bias. d. Expenditures of HMOs are directly related to the
d. Nonexperimental design and limited internal validity. quality of care they provide.
17. At 30 days after onset of a depressive episode, 24. Pharmacologic options that show considerable
nearly 75 percent of patients insured through a efficacy for SAD or generalized SAD include the
commercial health plan receive follow-up care, following classes of drugs with the exception of:
compared to approximately 60 percent through a. SSRIs.
Medicaid and Medicare. b. Reversible and irreversible MAO inhibitors.
a. True. c. Tricyclic antidepressants.
b. False. d. Benzodiazepines.
18. Even with antidepressant treatment, relapse of 25. Paroxetine is indicated for a wide range of anxiety
major depression occurs in ___ percent of and mood disorders, with the exception(s) of:
patients: a. Obsessive-compulsive disorder.
a. 15 percent. b. Premenstrual dysphoric disorder.
b. 10–30 percent. c. Social anxiety disorder.
c. 20–40 percent. d. Bulimia nervosa.
d. 50 percent. e. Answers a and d.
f. Answers b and d.

DEPRESSION 59
 CONTINUING EDUCATION ANSWER SHEET/CERTIFICATE REQUEST
P&T DIGEST Depression

CE Credit for Physicians/Pharmacists

Sponsored by Please allow up to 6 weeks for pro- 5. Elucidate strategies for therapeutic
The Chatham Institute cessing. compliance and the consequences
I certify that I have completed this of noncompliance. ■ Yes ■ No
educational activity and post-test and The cost of this activity is provided at
claim (please check one) no charge to the participant through 6. Discuss NCQA guidelines for depression-
medication management and follow-up
___ Physician credit hours an educational grant from Glaxo-
after hospitalization for depression.
___ Pharmacist contact hours SmithKline. ■ Yes ■ No
Signature: _______________________ EXAMINATION: Place an X through the box
of the letter that represents the best answer
to each question on pages 58 and 59. There Was this publication fair, balanced, and
First name, MI ____________________ free of commercial bias? ■ Yes ■ No
is only ONE answer per question. Place all
answers on this form:
Last name, degree ________________ If no, please explain: __________________
A. B. C. D. E. F.
1. ■ ■ ■ ■
Title ___________________________ 2. ■ ■ ■ ■ ___________________________________
3. ■ ■ ■ ■ ■
Affiliation _______________________ 4. ■ ■ ■ ■ ___________________________________
5. ■ ■ ■ ■ ■
Specialty ________________________ 6. ■ ■ ■ ■ ___________________________________
7. ■ ■
8. ■ ■ ■ ■ Did this educational activity meet my
Address ________________________ needs, contribute to my personal effec-
9. ■ ■
10. ■ ■ ■ ■ tiveness, and improve my ability to:
City__________ State ___ ZIP_____ Strongly Strongly
11. ■ ■ ■ ■
12. ■ ■ ■ ■ ■ agree disagree
Daytime telephone (____) __________ 13. ■ ■ ■ ■ ■ Treat/manage patients?
14. ■ ■ ■ ■ 5 4 3 2 1 N/A
Fax (_____) _________________ 15. ■ ■ ■ ■ ■ ■ Communicate with patients?
16. ■ ■ ■ ■ 5 4 3 2 1 N/A
E-mail __________________________ 17. ■ ■ Manage my medical practice?
18. ■ ■ ■ ■ 5 4 3 2 1 N/A
Physician — This activity is desig- 19. ■ ■ ■ ■
nated for a maximum of 3.0 category 1 20. ■ ■ ■ ■ Other _____________________________
credits toward AMA Physician’s Recog- 21. ■ ■ ___________________________________
22. ■ ■ 5 4 3 2 1 N/A
nition Award. 23. ■ ■ ■ ■
24. ■ ■ ■ ■ Effectiveness of this method
Pharmacists — This activity is ap- 25. ■ ■ ■ ■ ■ ■ of presentation:
proved for 3.0 contact hours (0.3 CEU). Very
PROGRAM EVALUATION Excellent good Good Fair Poor
ACPE Universal Program Number So that we may assess the value of this self- 5 4 3 2 1
(UPN): 812-999-04-009-H01 study program,please fill out this evaluation
Release date: June 15, 2004 form. Time spent reading this publication:
Expiration date: June 15, 2005
Have the activity’s objectives been met? Hours _____ Minutes _____
To receive CME credit, complete the 1. Illustrate the prevalence of depression,
answer sheet/evaluation form and as well as the clinical and economic What other topics would you like to see
mail or fax to: implications thereof. ■ Yes ■ No addressed? ________________________
Office of Continuing Education
The Chatham Institute 2. Describe major published treatment ___________________________________
26 Main Street, 3rd Floor guidelines for depression, as well as their
shortcomings as advanced therapies have ___________________________________
Chatham, NJ 07928 reached the market. ■ Yes ■ No
Fax: (973) 701-2515 Comments:_________________________
3. Identify various SSRI treatments for
Credit will be awarded upon success- depression and explain their mechanisms ___________________________________
ful completion of assessment ques- of action. ■ Yes ■ No
tions (70 percent or better) and com- 4. Understand the dynamics of pharmaco- ___________________________________
pletion of program evaluation. If a economic evaluation of SSRI therapy.
score of 70 percent or better is not ■ Yes ■ No ___________________________________
achieved, no credit will be awarded
and the registrant will be notified.

60 P&T DIGEST