Rebooting Psychotherapy Research
and Practice to Reduce the Burden
of Mental Illness
Alan E. Kazdin and Stacey L. Blase
Department of Psychology, Yale University, New Haven, CT
Abstract
Psychological interventions to treat mental health issues have developed remarkably in the past few decades. Yet this progress
often neglects a central goal—namely, to reduce the burden of mental illness and related conditions. The need for psychological
services is enormous, and only a small proportion of individua ls in need actually receive treatment. Individua l psychotherapy, the
dominant model of treatment delivery, is not likely to be able to meet this need. Despite advances , mental health professionals are
not likely to reduce the prevalence, incidence, and burden of mental illness without a major shift in intervention research and
clinical practice. A portfolio of models of delivery will be needed. We illustrate various models of delivery to convey
opportunities provided by technology, special settings and nontraditional service providers, self-help interventions, and the media.
Decreasing the burden of mental illness also will depend on integrating prevention and treatment, developing assessment and a
national database for monitoring mental illness and its burdens, considering contextual issues that influence delivery of treatment,
and addressing potential tensions within the mental health professions. Finally, opportunities for multidisciplinary collaborations
are discussed as key considerations for reducing the burden of mental illness.
Keywords
psychological interventions, reducing the burden of mental illness
Psychological interventions to treat clinical dysfunction have
advanced remarkably in the past few decades. The progress
is evident in many ways. First, the quantity of controlled treat-
ment outcome studies has proliferated. Empirical studies of
therapy for children, adolescents, and adults number well into
the thousands. Many journals feature therapy outcome research
as their primary thrust so the flow of research continues. Sec-
ond, the quality of research has continued to improve as well.
The use of randomized controlled trials (RCTs) is recognized
as the fundamental design, but many other methodological fea-
tures (e.g., the use of treatment manuals, assessment of clinical
significance of change, evaluation of follow-up) have set the
bar high for treatment outcome studies. Third, and perhaps
most salient, has been the delineation of evidence-based treat-
ments (EBTs; i.e., interventions with strong evidence on their
behalf). EBTs are available for many psychological dysfunc-
tions for children, adolescents, and adults (e.g., Nathan &
Gorman, 2007; Weisz & Kazdin, 2010). EBTs continue to
emerge and reflect palpable progress from scientific research.
The remarkable progress has left in the background a key
issue that is a major impetus for developing psychological
interventions—namely, the goalofdecreasingratesofmental
illness and improving psychosocial functioning on a large
scale (i.e., in society). Psychological treatments have many
purposes, but ke y among them is to alleviate mental illnesses
and related sources of dysfunction. A central thesis of this
article is tha t, despite a dvanc es in r esear ch, mental health pro-
fessionals may have little success in decreasing the preva-
lence and incidence of mental illness without a major shift
and expa nsion of intervention research and clinical prac tice.
The article focu ses on models of treatment delivery and what
is needed to reduce the burden. By bur de n,werefertotheper-
sonal, social, and mo netar y costsassociatedwithimpairment.
Within the term mental illness,weincludepsychiatricdisor-
ders and also social, cognitive, emotional, and behavioral
sources of impairment or disability.
Corresponding Author:
Alan E. Kazdin, Department of Psychology, 2 Hillhouse Avenue, Yale
University, New Haven, CT 06520-8205
E-mail: alan.kazdin@yale.edu
Perspectives on Psychological Science
6(1) 21–37
ª The Author(s) 2011
Reprints and permission:
sagepub.com/journalsPermissions.nav
DOI: 10.1177/1745691610393527
http://pps.sagepub.com
by guest on February 27, 2011pps.sagepub.comDownloaded from
We begin by highlighting the burden and cost of mental
illness and associated psychological sources of dysfunction.
We then discuss why advances in current treatment are not
likely to have broad impact and reach most people in need.
We highlight individual psychotherapy as a point of departure
because it serves as the dominant model of treatment delivery
and is emphasized in treatment research, clinical practice, and
training in the mental health professions. By model of delivery,
we refer to multiple characteristics of how an intervention is
administered, by whom, under what conditions, and in what
contexts. Psychotherapy as a model usually is delivered to one
person at a time (or couple, family, or in a small group) by a
trained mental health professional at a health care or mental
health service facility or private office. Although there are
many variations of therapy (techniques), the model of delivery
is more narrowly restricted among them. Yet emphasis on this
one delivery model leaves enormous gaps that must be
addressed to reduce the burden of mental illness.
In addition, this art icle highlights research on the burden of
mental illness and the current treatment model as steps toward
elaborating changes that are needed for providing treatment.
We illustrate several models of delivery that expand on the
model of individual psychotherapy. The burden of mental ill-
ness can be reduced by expanding models of delivery. At the
same time, reducing the burden raises other considerations
including the integration of prevention and treatment, the need
for improved assessment to monitor psychological dysfunction
nationwide, contextual factors that influence health care, pro-
fessional tensions within clinical psychology, and important
opportunities for collaborating with other disciplines.
Reducing the Burden of Mental Illness and
Related Conditions
The challenge for psychological interventions is to help reduce
the burden of mental illne ss and related conditions both at the
personal and societal level.
1
Four interrelated considerations
convey why diverse treatment delivery models are needed.
Rate of Mental Illness
Consider the rate of psychological dysfunction. Not all
sources of psychological impairment are codified by current
classification systems of psychiatric disorders (Diagnostic and
Statistical Manual of Mental Disorders, DSM-IV-TR;Ameri-
can Psychiatric Association, 2000; International Classification
of Diseases, ICD-10; World Health Organization, 2007).
Impairment can result from many sources (e.g., stress, relation-
ship problems) beyond those included in diagnostic systems.
Also, many disorders are on a spectrum indicating continuity
of dysfunction. For example, several symptoms of depression
are required to meet criteria for a DSM diagnosis. However,
just failing to meet the criteria (e.g., by a symptom or duration
requirement) is still associated with dysfunction or impairment,
commensurate with the discrepancy from meeting the criteria.
Subthreshold dysfunction leads to prevalence rates that
underestimate the burden of dysfunction. Nevertheless, data
on psychiatric disorders, albeit conservative, are instructive
in illustr ating the scope of psychological dysfunction.
Lifetime preval ence rates reveal that mental illness (meeting
criteria for a psychiatric disorder) is relatively common, not
only within the United States, but also within many countries
worldwide. A series of recent surveys from the World Health
Organization assessing the global burden of mental illness
found a lifetime DSM-IV disorder prevalence within its 17 par-
ticipating countries of 12.0% to 47.4%, with the highest life-
time prevalence estimate in the United States (Kess ler et al.,
2009). The same surveys reported the United States to have the
highest 12-month DSM-IV disorder prevalence, with a range of
6.0% to 27.0% for all 17 countries. Summarizing the U.S. data
only, approximately 50% of the population meets criteria for
one or more psychiatric disorders in their lifetimes, and approx-
imately 25% of the population meets criteria in any given year
(Kessler & Wang, 2008).
The rates of dysfunction vary as a function of culture, ethni-
city, immigrant status within a given ethnicity, geographical
location, and socioeconomic status, among ot her factors (e.g.,
Alegrı´a et al., 2008). These variations and differences are
important in developing interventions. For present purposes,
we merely wish to convey that psychiatric disorders are preva-
lent. The estimates are likely to be conservative because they
have required meeting diagnostic criteria and exclude those
who do not quite meet criteria but are close enough to make the
distinction of meeting or not meeting the criteria minor.
Cost of Mental Illness
The costs of mental il lness are high. Although there is no sin-
gle, agree d up on f ig u re o r s et o f fi gu r es o f th os e costs, we ll -
documen te d exampl e s in spec if ic pro b le m doma in s conv ey
the point. For exampl e, alcoholism and substa nc e a bu se,
which affects more than 2 0 million Americans and is the most
prevalent mental disorder in the United States, costs approx-
imately $500 bill ion annually (Jason & Ferrari, 2010). The
main costs inc lude me di cal a nd criminal justice costs, acci-
dents, and loss o f earnin g s. For a nxiet y dis orde rs , annu al
health-care expenditures in the United States are approxi-
mately $42 billion (Greenberg et al., 1999). The costs encompass
health-care utilization, including medical and psychiatric treat-
ment, and decreased work productivity (see also H. Harwood,
Fountain, & Livermore, 1998).
Reductions in annual earnings also are associated with the
diagnosis of a mental illness. Individuals diagnosed with a
DSM-IV mental disorder earn, on average, approximately
$16,000 less than their control counterparts annually. This results
in a total reduction of $193.2 billion in personal earnings nation-
ally in 1 year (Kessler et al., 2008). A single episode of major
depressive disorder is associated with an average of more than
5 weeks of lost productivity per worker, resulting in an annual
capital loss of $36 billion to employers (Kessler et al., 2006).
Cost extends beyond the fiscal burden; personal impairment
and subjective experience are not trivial. In one series of
22 Kazdin and Blase
by guest on February 27, 2011pps.sagepub.comDownloaded from
national interviews, participants reported the number of days in
the past month in which they were unable to perform th eir usual
daily tasks due to problems because of physical or emotional
health. Mental disorders were associated with more than half
of the role-disability days, and depression had one of the largest
effects on disability of all conditions (Merikangas et al., 2007).
Mental disorders are more impairing than common chronic
medical disorders, with particularly greater impairment in the
domains of home, social, and close relationship functioning
(Druss et al., 2009). These findings document the importance
of the loss of productivity due to mental illness and also reflect
the widespread interpersonal difficulties. These psychosocial
sources of dysfunction that likewise lead to impairment, suffer-
ing, and costs to individuals as well as to society are more dif-
ficult to capture on a national level, but they are an important
target for interventions to reduce the burden of mental illness.
People in Need of Services
Avitalaspecttoreducingtheburdenofmentalillnessisthe
availability of interventions for those most in need of services.
Recent years have seen an increase in the rate of people in need
receiving treatment in the United States, with 20.3% of individ-
uals suffering from a disorder receiving treatment between
1990 and 1992 and 32.9% receiving treatment between 2001
and 2003 (no difference was found in rates of prevalence
between the two tim e points; Kessler et al., 2005). Even so, the
majority of individuals with a diagnosable mental disorder are
not receiving treatment. Ethnic disparities with respect to
access to mental-health care among those in need are enor-
mous. For example, African Americans are less likely to have
access to services than are European Americans (12.5 vs.
25.4%), and Hispanic Americans are less likely to have ade-
quate care than are European Americans (10.7 versus 22.7%;
Wells, Klap, Koike, & Sherbourne, 2001). These are illustra-
tions from a much larger literature on disparities in mental
health care delivery among individuals of minority groups
(e.g., see www.mentalhealthcommission.gov/reports/Final
Report/FullReport-04.htm). The lack of available services for
most people and systematic disparities among those services
have direct implications for models of treatment delivery.
Interventions are needed that can reach many more people, but
also with particular attention to select subpopulations.
People Providing Services
The problem of too few people receiving services and of groups
being particularly deprived of such services might be con-
ceived as a ‘‘person-power problem’’ in the mental health field.
To oversimplify the argument, maybe more people are needed
to provide the usual services. Three points convey why more
providers a lone may not be sufficient. First, the person-
power problem stems in part from the geographical distribu-
tion of existing mental health professionals. The concentra-
tion of psychologis ts , psyc hi atr ists, an d clinic al soc ial
workers in the United Statesisthegreatestinhighly
populated, afflu ent ur ba n area s and in ci ties wi th maj or
universities (Health Resources and Services Administration,
2010; J.M. Richards & Gottfredson, 1978). Thus, mental health
professionals are not distributed advantageously to reach large
swaths of people in need (e.g., rural areas, small towns).
Second, mental health profe ssionals are unlikely to be able
to reach the vast majority in need. In the United States, there
are approximately 700,000 mental health professionals who
provide services (Hoge et al ., 2007). As we ment ion ed, esti-
mates of prevalence indic ate that approximately 25% of the
population has at le as t one psychiatri c disorder in a given
year. With a U.S. population exceeding 300 million; 25% is
approximately 75 million people. It is not necessary for these
estimates to be precise to see that the number of individuals
who can pr ovide treatment—at le ast with the current domi-
nant model of delive ry— c ou ld not begin to reach the numbe r
of those in need. Doubling the work force might have little
discerni b l e impa ct give n the number of individ uals requiring
services.
Third, it is not only the distribution and number but the pro-
file of mental health professionals. Too few mental health pro-
fessionals reflect the cultural and ethnic characteristics of those
in need of care. For example, trends in ethnic minority repre-
sentation in clinical psychology (e.g., doctoral enrollments and
recipients and graduate department faculty) over two decades
suggest that the proportions of ethnic minorities in psychology
do not show comparable growth relative to that of the U.S.
ethnic minority population growth (e.g., see Commission on
Ethnic Minority Recruitment, Retention, and Training, 2008).
Approximately 20% of doctoral degrees in psychology and cur-
rent enrollees in psychology graduate training programs are of
minority status, and approximately 6% of psychology faculty
are of minority status (African American, Asian American,
Hispanic American, Native American). The proportion of the
U.S. population that comprises minorities, projected to be
50% by 2060, is accelerating at a higher rate than trainees in
their respective groups. Thus, the population of ethnic minori-
ties in need of mental health services will increase at a greater
rate than will the availability of ethnically matching profes-
sionals. Although treatments can be effective where there is not
an ethnic match of therapist and client, in some cases that cul-
tural component influences outcome (Griner & Smith, 2006;
Miranda et al., 2005). Even if an ethnic match is not needed for
treatment to work, it can nevertheless present a barrier for the
potential client who is considering treatment. Ethnicity and cul-
ture are not the only mismatch. Other groups based on geogra-
phy (e.g., individuals living in rural areas) or select populations
(e.g., the elderly) reflect a mismatch with a paucity of available
resources and a plethora of need for mental health services
(e.g., Hinrichsen, 2010; Institute of Medicine, 2008).
In light of the previous considerations, the inability to reach
most people in need of services is not simply (or only) a person-
power issue. Many of those in need of services cannot be
reached for a variety of reasons (e.g., access, perceived and
genuine barriers in obtaining treatment, insurance, rural areas),
but one of them is our own view as mental health professionals
Rebooting Psychotherapy Research and Practice 23
by guest on February 27, 2011pps.sagepub.comDownloaded from
regarding how treatment should be delivered. Invariably, more
help is welcome. But, given the emphases of the current model
of delivery, that alone is not likely to provide an increment of
reduction on the burden of mental illness.
Individual Psychotherapy in Relation to the
Burden
Many interventions will be needed to reduce the burden of
mental illness and other facets of impairment. We begin with
psychotherapy research for several reasons.
2
First, there are
many goals of psychotherapy, but salient among them is the
treatment or amelioration of psychiatric disorders; social, emo-
tional, cognitive, and behavioral problems; and stress (e.g.,
Dickerson & Lehman, 2006; Mahrer, 1986; Weissman,
Markowitz, & Klerman, 2000). Thus, psychotherapy is a viable
intervention for addressing significant mental health problems
(e.g., anxiety, depression, bipolar disorder) in addition to other
sources of impairment (e.g., stress). Second, psychotherapy
research has progressed remarkably. As we mentioned,
the emergence of EBTs is a particularly important advance.
Generally, EBTs refer to those interventions that have carefully
controlled research on their behalf. RCTs, careful delineation
of the client sample, specification of treatment, and replication
of the results by an independent investigator or team are among
the commonly used criteria. Third, psychotherapy is a domi-
nant model of delivering psychological services. By model of
delivery, we refer to how the intervention is presented or pro-
vided. In this case, we refer to the dominance of individual,
one-to-one therapy with a client (child or adult), family, or
group (e.g., 8–10 clients). Psychotherapy plays a role in reduc-
ing the personal and social burden of mental illness. But the
role it does play draws stark attention to what is missing if the
burden of dysfunction is to be significantly reduced among
those in need.
It is useful to highlight briefly the model of delivery of psy-
chotherapy in historical context. That context conveys how the
model has tacitly continued to dominate in therapy research
and practice. Consider the enormous impact of psychoanalysis
on the delivery of psychological treatment. For present pur-
poses, three facets of psychoanalysis and its variations can be
distinguished. First is the theory or substantive foci of problem
development (e.g., related to psychosexual stages of develop-
ment, superego, and variants) and of treatment (e.g., addressing
transference). Second are the methodological features used in
early work to support the key tenets. The anecdotal case study
(e.g., Anna O., Dora, Little Hans) was relied on heavily. Third
was the model of delivery of treatment, namely, one-to-one
individual patient care, all flowing from a medical-patient care
model. Psychoanalysis was not the first one-to-one psychoso-
cial intervention (e.g., mesmerism, hypnotism), but it provided
aprototypefromwhichsubsequentpsychotherapies(andNew
Yorker cartoons of them) followed.
The contemporary history of psychotherapy research
reflects the continual development and changes in substantive
views that ex plain the o nset of clinical disorders and the
procedures required to effect therapeutic change. Many
familiar examples within psychoanalysis and its variants (e.g.,
Jung, Kohut) and also from other views (e.g., self-theory
of Rogers, reciprocal inhibition of Wolpe, cognitive therapy of
Beck) and shifts in orientation (e.g., positive psychology of
Seligman) illustrate the dynamic (not ‘‘psycho’’) nature of the
field. Also, methodological advances have raised the bar for all
treatment evaluation to include RCTs and increasingly more
stringent requirements for how intervention studies must be con-
ducted and reported (e.g., Consolidated Standards of Reporting
Trials—CONSORT; Moher, Schulz, & Altman, 2001).
3
The third component, the model of one -to-one treatment,
continues to dominate even as theories about the appropriate
intervention focus (e.g., problem-solving strategies, mindful-
ness, and self-agency) have proliferated. The departures (e.g.,
treating couples, families) retain the focus on small individual
units. The majority of EBTs retain this model of delivering
treatment (Nathan & Gorman, 2007; Weisz & Kazdin, 2010).
Also, the model dominates training in clinical psychology,
social work, and psychiatry. In clinical psychology, for exam-
ple, accreditation of graduate programs emphasizes clinical
hours of tre atment in individual therapy in graduate school fol-
lowed by an internship experience. Treatment of groups (e.g.,
8–10) and families are counted as well, but in relation to the
focus of this article, delivering treatment by a mental health
professional in person to small individual units are of the same
ilk, namely, treatments with a very restricted reach.
Our comments are not a criticism of the model of individual
therapy. One-to-one therapy will always have a place; individ-
ual crises and challenges in life are invariably at that level.
Also, individual therapy contributes to the overall goal of
reducing the burden of mental illness in at least three ways.
First, it serves a small number of individuals with effective pro-
cedures, and that places it firmly in the portfolio of models of
treatment delivery. Second, many of the scientific principles
and processes (e.g., emotional memories, extinction, cognition)
that serve as the underpinnings of individual therapy as well as
the techniques themselves may inform other models of deliv-
ery. Third, some therapy techniques (as noted later) might be
delivered in multiple ways, so the ‘‘same’’ treatment or very
close approximations may vary in their accessibility and reach.
Nevertheless, additional delivery models will be needed
beyond the contributions of individual therapy.
Developing a Portfolio of Models of
Delivering Interventions
Interventions that can reduce prevalence (cases with some dys-
function now) and incidence (new cases that emerge) are
needed to reduce the burden of mental illness. Treatment and
prevention work arm in arm. We emphasize treatment to con-
vey key points and return to its integration with prevention later
in this article.
Consider all individuals in need of psychological services
(treatment) as occupying a pie chart. The goal in developing
multiple models of delivery is to ensure that segments (slices)
24 Kazdin and Blase
by guest on February 27, 2011pps.sagepub.comDownloaded from
of the pie are covered (i.e., services encompass all or most in
need—at least in principle). One might consider slices of the
pie with the view that a given intervention or model of delivery
may reach one slice, but that multiple models might be needed
to cover more, most, or all of the slices. The pie notion is use-
ful, but it does not convey the multidimensional needs of the
population. There are many different reasons why people do
not receive services, such as lack of access to facilities or prac-
titioners, ethnic and cultural barriers, and many concrete obsta-
cles (e.g., transportation, babysitting). No single model of
delivery can be assured to circumvent all of the obstacles asso-
ciated with a given subgroup or slice of the population pie.
Among the many characteristics that might delineate models
of delivery, the ability to reach many individuals in need of ser-
vices reflects the type of changes that are needed if treatment is
to significantly reduce the burden of mental illness. We illustrate
models of delivery of psychosocial interventions to convey some
of the many options that might comprise the portfolio.
Technologies
The Internet. Unlike individual therapy, the use of various
technologies to deliver treatment has the abilit y to reach a large
proportion of the population in need of services. Among the
technological options, there is a rapidly growing literature on
the use of the Internet (e.g., Barak, Hen, Boniel-Nissim, &
Shapira, 2008; Dimeff, Paves, Skutch, & Woodcock, 2011;
Ybarra & Eaton, 2005). The ability to reach a large segment
of the population in need is nicely illustrated in an application
to cigarette smoking, which is often a target of psychological
interventions.
AseriesofWeb-basedinterventionstudiesforsmoking
cessation conducted in English and Spanish have shown signif-
icant smoking termination rates through a standard smoking
cessation guide and mood management course (Mun
˜
oz et al.,
2006). An individualized, password-protected Web site pro-
vided access to the smoking cessation intervention to conse nt-
ing eligible individuals and was used to obtain assessment data
throughout the intervention. The interventi on reached more
than 4,000 smokers from 74 countries and was carefully eval-
uated (e.g., RCT, follow-up assessments). Studies of this pro-
gram illustrate the potential for use of Web-based
interventions and the ability to reach people in their homes.
Once developed, such Internet-based administrations can be
relatively inexpensive to implement and easy to maintain.
Telephone. As with the Internet, the telephone can be used to
deliver mental health interventions for both individuals and
groups (see Mohr, Vella, Hart, Heckman, & Simon, 2008). Typi-
cally, ‘‘telemental psychotherapy,’’ as this is sometimes called,
involves administration of full sessions of therapy through
scheduled weekly phone calls. This is much like the model of
in-person psychotherapy, but because no face-to-face contact
is necessary, it allows for remote administration of services.
Interestingly, telephone-administered psychotherapies have
lower rates of attrition than traditional individual psychotherapy
(Mohr et al., 2008). Thus, phone-based treatment may not only
broaden the population with access to therapeutic intervention,
but also p otentially increase the likelihood that clients will
remain in treatment. Such a model has low cost and may even
reach population segments to which Internet-delivered models
may not have access.
One such illustration of the pot ential for telephone-based
intervention is nicely demonstrated in ‘‘quitlines,’’ initially
developed to provide telephone counseling for smoking
cessation through th e U.S. National Cancer Institute. Tobacco
users who call a quitline receive an empirically validated, stan-
dardized, and manualized intervention incorporating various
services such as materials by mail, prerecorded messages,
real-time phone counseling or a return phone call from a coun-
selor, access to quitting medication, or some combination
thereof (Lichtenstein, Zhu, & Tedeschi, 2010). Quitlines have
demonstrated tremendous reach, as they are currently offered
in all 50 states and Washington, DC and have also been adopted
and sponsored at the national level of various countries in
Europe, Oceania, Asia, and South and North America.
Although the content and structure vary across quitlines, the
initial call typically lasts less than an hour, orients the client
to the program, and establishes a quit date. Subsequent calls
of 10 to 15 min are scheduled following the quit date over a
period of 1 or more months with a frequency based on a relapse
curve (Zhu & Pierce, 1995). A staff with bachelor’s- or master’s-
leve l training typically delivers the counseling services, although
computers drive much of the quitline counseling. An estimated
1% of smokers in the United States utilize quitline services each
year (Cummins, Bailey, Campbell, Koon-Kirby, & Zhu, 2007),
and some states with increased marketing reach as many as 4%
to 5% of their smoking populations in a single year (Swartz
Woods & Haskins, 2007). Quitlines have even demonstrated a
special ability to reach underserved populations, as African
American smokers are more likely than any other ethnic group
to utilize these services and Asian immigrant smokers are as
likely as European American smokers to utilize them (Maher
et al., 2007; Zhu, Wong, Stevens, Nakashima, & Gamst,
2010). This program highlights the advantages of models utilizing
telephone-based intervention, namely, the potential to overcome
various logistical barriers to treatment that exist for in-person
individual psychotherapy. This and the brief, standardized,
semianonymous nature of pho ne counseli ng administered by
paraprofessionals greatly increases the accessibility and reach
of the i nterv e ntion .
Smartphones. Due to advances in technology, cellular phones
are no longer simply mobile telephones for the sole purpose of
making calls. Updates in these devices (e.g., GPS) provide new
opportunities for methods of intervention and assessment (e.g.,
loca tion of client engaged in homework assignments to over-
come fear of open spaces; Boschen, 2009a, 2009b). The most
commonly studied form of mobile phone intervention employs
the use of short message service (SMS) or text messaging. One
example of implementing an SMS-based intervention is nicely
illustrated in an aftercare treatment of bulimia nervosa. Patients
received weekly text messages for the 6 months following their
release from inpatient psychotherapy. These messages provided
Rebooting Psychotherapy Research and Practice 25
by guest on February 27, 2011pps.sagepub.comDownloaded from