Gregory E. Simon, Evette J. Ludman, Steve Tutty, Belinda Operskalski, Michael Von Korff. Telephone Psychotherapy and Telephone Care Management for Primary Care Patients Starting Antidepressant TreatmentA Randomized Controlled Trial. JAMA. 2004;292(8):935–942. doi:10.1001/jama.292.8.935
Author Affiliations: Center for Health Studies, Group Health Cooperative, Seattle, Wash.
Context Both antidepressant medication and structured psychotherapy have been
proven efficacious, but less than one third of people with depressive disorders
receive effective levels of either treatment.
Objective To compare usual primary care for depression with 2 intervention programs:
telephone care management and telephone care management plus telephone psychotherapy.
Design Three-group randomized controlled trial with allocation concealment
and blinded outcome assessment conducted between November 2000 and May 2002.
Setting and Participants A total of 600 patients beginning antidepressant treatment for depression
were systematically sampled from 7 group-model primary care clinics; patients
already receiving psychotherapy were excluded.
Interventions Usual primary care; usual care plus a telephone care management program
including at least 3 outreach calls, feedback to the treating physician, and
care coordination; usual care plus care management integrated with a structured
8-session cognitive-behavioral psychotherapy program delivered by telephone.
Main Outcome Measures Blinded telephone interviews at 6 weeks, 3 months, and 6 months assessed
depression severity (Hopkins Symptom Checklist Depression Scale and the Patient
Health Questionnaire), patient-rated improvement, and satisfaction with treatment.
Computerized administrative data examined use of antidepressant medication
and outpatient visits.
Results Treatment participation rates were 97% for telephone care management
and 93% for telephone care management plus psychotherapy. Compared with usual
care, the telephone psychotherapy intervention led to lower mean Hopkins Symptom
Checklist Depression Scale depression scores (P =
.02), a higher proportion of patients reporting that depression was "much
improved" (80% vs 55%, P<.001), and a higher proportion
of patients "very satisfied" with depression treatment (59% vs 29%, P<.001). The telephone care management program had smaller
effects on patient-rated improvement (66% vs 55%, P =
.04) and satisfaction (47% vs 29%, P = .001); effects
on mean depression scores were not statistically significant.
Conclusions For primary care patients beginning antidepressant treatment, a telephone
program integrating care management and structured cognitive-behavioral psychotherapy
can significantly improve satisfaction and clinical outcomes. These findings
suggest a new public health model of psychotherapy for depression including
active outreach and vigorous efforts to improve access to and motivation for
Abundant evidence supports the efficacy of antidepressant pharmacotherapy
and structured psychotherapy for treatment of depression.1- 3 Unfortunately,
the reach and actual effectiveness of either treatment remain poor. Of patients
who begin taking antidepressants, 40% discontinue within a month, and only
25% receive even minimal levels of follow-up.4,5 Only
one third of patients with depressive disorders receive any psychotherapy.6 Of those beginning psychotherapy, 25% attend only
1 session and only half attend 4 or more sessions.7 At
the population level, only 25% to 30% of those with depressive disorders receive
an effective level of either treatment.6,8
While shortcomings in the care of depression parallel those seen for
other chronic illnesses,9,10 depression
treatment presents special challenges. Depression is defined by pessimism,
discouragement, and fatigue. Stigma remains an important barrier to treatment
seeking and treatment adherence.11 Benefits
of treatment may not appear for several weeks. Antidepressant pharmacotherapy
requires persistence through adverse effects and medication adjustments.12 Psychotherapy requires a significant commitment of
time. Dissemination of empirically supported psychotherapies is slower and
more difficult than dissemination of new medications or medical procedures.13
We describe a randomized trial evaluating 2 approaches to addressing
these barriers to effective depression treatment. The first program was an
updated version of a telephone outreach and care management program to improve
the quality of antidepressant pharmacotherapy.14 The
second program included telephone care management and added an 8-session structured
psychotherapy program delivered by telephone.15 Both
were compared with usual primary care treatment.
Between November 2000 and May 2002, participants were sampled from 7
group-model primary care clinics of Group Health Cooperative (GHC). The GHC
is a prepaid health plan serving approximately 500 000 Washington State
residents enrolled via employer and individual contracts and via risk-sharing
contracts with Medicare and Medicaid programs. The GHC membership is demographically
similar to the Seattle area population.16
Procedures were designed to enroll a population-based sample of primary
care patients beginning antidepressant treatment for depression, excluding
those already receiving psychotherapy and those already in remission when
contacted. Each week, computerized pharmacy and visit registration data identified
all adult patients starting a new episode of antidepressant medication for
treatment of depression (ie, filled antidepressant prescription, visit diagnosis
of depression, and no antidepressant use in the prior 90 days). Computerized
records were also used to exclude those already receiving psychotherapy (ie,
any specialty mental health visit in the last 90 days) and those with a diagnosis
of bipolar disorder or schizophrenia in the last 2 years.
Potential participants received an invitation letter including all elements
of informed consent (full description of study procedures, risks, and potential
benefits). A telephone interview approximately 1 week later assessed eligibility
and baseline depression severity (Hopkins Symptom Checklist17 Depression
Scale [SCL] and Patient Health Questionnaire18 [PHQ]).
Those already in remission at the baseline assessment (ie, SCL depression
score <0.5) were excluded. Interviewers also asked about use of non-GHC
services to exclude those not meeting primary eligibility criteria (ie, not
a new episode of antidepressant treatment, receiving or planning to receive
psychotherapy). Those with cognitive, language, or hearing impairment severe
enough to preclude participation were also excluded.
Eligible patients were invited to participate in the randomized trial.
The 2 interventions were briefly described, but commitment to participate
in either intervention was not a condition for randomization. After a full
description of study procedures, risks, and benefits, all participants provided
documented oral consent prior to the baseline assessment and again prior to
enrollment in the randomized trial. All study procedures, including use of
oral consent, were reviewed and approved by the GHC Human Subjects Review
Between 1 and 7 days after the baseline interview, the study data manager
assigned eligible and consenting participants to 1 of 3 treatment groups using
computer-generated random numbers without blocking or stratification. Participants
assigned to either of the 2 intervention programs were notified by a telephone
call from the care manager or telephone counselor (see below). Participants
assigned to the usual care control group were not contacted until the first
blinded outcome assessment (see below).
Care managers contacted participants assigned to the program within
2 weeks of randomization (typically 4 weeks after the initial antidepressant
prescription). Two additional telephone contacts occurred 4 and 12 weeks later
with a personalized mail contact approximately 20 weeks later. As in our earlier
care management program,14 each contact included
a brief, structured assessment of depressive symptoms, antidepressant medication
use, and adverse effects. During telephone contacts, care managers followed
specific scripts to address concerns regarding adverse effects and used scripted
motivational enhancement interventions to address common reasons for discontinuing
The treating primary care physician received a structured report of
each contact including a summary of the clinical assessment and computer-generated
recommendations regarding medication adjustment. If a change in treatment
was recommended, the care manager contacted the treating physician to facilitate
patient-physician communication and follow-up. Care managers also provided
as-needed crisis intervention and care coordination including referrals to
mental health specialty care. Most care manager time was spent in outreach
(eg, ≥5 telephone attempts to complete a contact, mailings to patients
not responding to telephone calls) and coordination with treating physicians.
After enrollment in the program, each participant received a detailed
self-management workbook (adapted from the telephone therapy patient workbook
described below) emphasizing behavioral activation, identifying and challenging
negative thoughts, and developing a long-term self-care plan. Care managers
recommended reading the workbook but did not provide any specific counseling.
All care management activities (caseload tracking, structured assessment,
medication algorithms, provider reports, supervision) were organized and supported
by an electronic decision support system.
Care managers (K.H. and S.H.) were mental health clinicians with bachelor's
or master's degrees and at least 1 year of experience in depression assessment
(including telephone assessment and triage). Additional training for this
study included 6 hours of didactic instruction and role-play followed by completion
of at least 5 observed care manager contacts prior to any patient contact.
Care managers received approximately 30 minutes of supervision each week from
a psychiatrist (G.E.S.) and psychologist (E.J.L.).
The telephone psychotherapy intervention included all aspects of the
telephone care management program plus a structured 8-session cognitive-behavioral
psychotherapy program adapted from an earlier pilot study.15 The
initial session occurred 1 to 2 weeks after randomization with sessions 2
through 4 at approximately weekly intervals. Intervals between later contacts
ranged from 1 to 4 weeks depending on need and patient preference. As with
the telephone care management program, each session began with a brief structured
assessment of depressive symptoms, medication use, and adverse effects. The
telephone psychotherapy program was an addition to, rather than a substitute
for, telephone care management. Therapy sessions were designed for completion
in 30 to 40 minutes. The initial session included a detailed clinical history,
assessment of motivation for treatment, and motivational enhancement exercises.19 Sessions 2 through 4 focused on increasing pleasant
and rewarding activities.20 Sessions 5 through
7 focused on identifying, challenging, and distancing from negative thoughts.21,22 Session 8 focused on creation of
a personal self-care plan covering medication use, self-monitoring, and self-management
A participant workbook included didactic material, in-session exercises,
and written homework exercises for completion between sessions. Participants
were asked to read the relevant workbook chapter prior to each session. During
each session, therapists followed a specific agenda and completed a detailed
checklist to monitor session content. Following each session, the therapist
mailed a personalized follow-up letter describing mutually agreed-on plans
for between-session homework. While therapists attempted to schedule sessions
in advance, failure to keep appointments was common and significant outreach
(eg, ≥5 outreach calls to complete a single session) was sometimes required.
Telephone counselors (S.T. and K.L.) were psychotherapists with master's
degrees and at least 1 year of experience in outpatient psychotherapy of depression.
Additional training included approximately 12 hours of didactic instruction
and role-play, trainee's observation of at least 6 sessions, and conduct of
6 sessions with audiotape observation. The 2 therapists received approximately
60 minutes of supervision each week from a psychologist (E.J.L.) and psychiatrist
(G.E.S.) and attended a twice-monthly seminar (led by E.J.L.) on motivational
interviewing techniques.24 Each therapist audiotaped
approximately 15 additional sessions for detailed review.
All participants were contacted for blinded telephone outcome assessments
6 weeks, 3 months, and 6 months after randomization. All assessments included
a patient-rated measure of global improvement since entering the study (a
7-point scale ranging from "very much improved" to "very much worse") and
repeat administration of the SCL depression scale. The 3- and 6-month assessments
also included the PHQ and a rating of satisfaction with depression treatment
on a 7-point scale ranging from "very satisfied" to "very dissatisfied." To
preserve blinding of telephone interviewers, participants were repeatedly
advised to offer no information regarding treatment received. Interviewers
had at least 1 year of experience conducting similar assessments.14,25 Training for this study included
6 hours of didactic instruction and role-play followed by weekly supervision
by the principal investigator. Previous research supports the reliability
and convergent validity of depression assessments conducted via telephone.26
Computerized records identified outpatient visits to primary care or
specialty mental health providers. Computerized pharmacy records were used
to compute the proportion of patients using antidepressant medication for
90 days or more at a minimally adequate dose (eg, 75 mg/d of imipramine or
10 mg/d of fluoxetine).27
Analyses used all available data at each time point with comparisons
based on original treatment assignment, regardless of treatment received.
Primary analyses (defined a priori) compared each intervention group with
the usual care group in terms of mean SCL depression score averaged across
the 3 follow-up assessments and in terms of 3 categorical outcomes at 6 months:
proportion of participants with at least a 50% improvement in SCL depression
score, proportion with self-ratings of "much improved" or "very much improved,"
and proportion with a rating of "very satisfied" with depression treatment.
Unadjusted comparisons used t tests for continuous
measures and χ2 statistics for categorical measures. Comparisons
of mean SCL depression scores over time used generalized estimating equations
with each follow-up assessment as a repeated measure and baseline depression
score, age, and sex as covariates. Power calculations indicated that a sample
of 200 per group was necessary to detect a difference of 0.2 in mean SCL depression
score (80% power, 2-sided P value = .05). All analyses
were conducted using version 8 of the SAS software package (SAS Institute,
The progress of participants through the trial is shown in Figure 1. Of 1883 sampled from computerized
records, 1247 (66%) participated in the telephone eligibility assessment.
Participants and nonparticipants did not differ in mean age, proportion female,
or prior history of depression treatment. Of 634 found eligible for randomization,
600 (95%) agreed to participate. Baseline characteristics of participants
assigned to the 3 treatment groups are compared in Table 1. On average, participants reported a moderate level of depressive
symptoms at baseline (2-4 weeks after starting treatment).
Of patients randomized, 578 (96%) completed at least 1 blinded follow-up
assessment and were included in analyses of clinical outcomes; 532 (89%) completed
the 6-month blinded assessment. Those completing and not completing the 6-month
assessment did not differ significantly in age, sex, baseline depression score,
or treatment assignment. Of those randomized, 563 (94%) remained enrolled
in the health plan for 6 months and were included in analyses of antidepressant
use and visit rates (utilization analyses).
Of those assigned to telephone care management, 97% completed at least
1 telephone contact and 85% completed all 3 contacts. Contacts typically lasted
10 to 15 minutes.
Of those assigned to telephone psychotherapy, 14 (7%) completed no sessions,
2 (1%) completed only the first session (history and motivational enhancement),
167 (84%) completed 4 or more sessions (including behavioral activation),
and 125 (63%) completed 7 or more sessions (including behavioral activation
and cognitive techniques). Sessions typically lasted 25 to 35 minutes, but
varied from 10 to 75 minutes.
As shown in Figure 2, improvement
in SCL depression score was greatest in the telephone psychotherapy group,
intermediate in the telephone care management group, and least in the usual
care group. In a repeated measures model, the telephone psychotherapy group
showed significantly lower mean depression scores during follow-up (χ21 = 5.94, P = .02) with an increasing
difference from 6 weeks to 6 months (P<.001).
The difference between the telephone psychotherapy and usual care groups at
6 months (0.33 points on the SCL depression scale) is equivalent to approximately
one half of the SD of scores in the general population. The mean difference
in follow-up depression scores between the telephone care management group
and usual care was not statistically significant (χ21 = 0.69, P = .40). Results for PHQ depression
scores closely paralleled those for SCL scores (data not shown). As shown
in Table 2, patients assigned
to telephone psychotherapy were significantly more likely to experience a
50% improvement in SCL depression score than were usual care patients. By
this definition, the number needed to treat per additional treatment response
The telephone care management group showed an intermediate rate of treatment
response by this measure, not significantly different from usual care. Both
telephone psychotherapy and telephone care management participants were significantly
more likely than usual care participants to describe themselves as "much improved"
or "very much improved." By this definition, number needed to treat for the
telephone psychotherapy program to achieve an additional treatment response
was 4.1. The pattern of results did not change if participants not completing
the 6-month assessment were considered nonresponders on both measures. The
proportion of patients "very satisfied" with treatment for depression was
significantly higher in both the telephone psychotherapy and telephone care
Secondary analyses compared clinical effects of the telephone psychotherapy
and telephone care management programs. Patients assigned to telephone psychotherapy
were significantly more likely to describe themselves as "much improved" or
"very much improved" (P = .004) and to be "very satisfied"
with depression treatment (P = .02). The 2 interventions
did not differ significantly in mean SCL depression score during follow-up
(P = .09) or in probability of 50% or greater improvement
in SCL depression score (P = .18).
As shown in Table 3, there
were modest differences in some categories of outpatient visits (eg, more
depression-related primary care visits in the telephone care management group,
fewer in-person psychotherapy visits in the telephone psychotherapy group).
Total visit rates, however, were generally similar. When in-person visits
were considered along with telephone psychotherapy visits, the proportion
of participants receiving at least 4 sessions of psychotherapy was, as expected,
markedly higher in the telephone psychotherapy group. The proportion receiving
at least 4 psychotherapy visits was higher in the telephone care management
group than in usual care, but rates in both groups were below 10%. Patients
assigned to telephone care management were significantly more likely to use
antidepressants at an adequate dose for at least 90 days. There was a roughly
similar increase in probability of adequate pharmacotherapy for the telephone
psychotherapy group, but the difference from usual care was not significant
at the 5% level.
Post-hoc analyses examined baseline factors predicting benefit from
the intervention programs. Effects of either intervention did not vary according
to participant age, sex, race/ethnicity, educational level, or marital status.
Intervention effect did vary according to baseline depression severity with
no apparent effect of either intervention among those with mild baseline SCL
depression scores between 0.5 and 1.0. Intervention effects were generally
similar for those with moderate and severe symptoms.
We found that a program combining telephone care management and brief,
structured psychotherapy significantly improved outcomes for primary care
patients initiating antidepressant treatment. Outcomes for the less intensive
telephone care management program were intermediate between the telephone
psychotherapy program and usual care. Our findings align with several other
randomized trials demonstrating the benefits of systematic care improvement
programs for depressed primary care patients.14,25,28- 34
Telephone programs may sacrifice the richness of traditional in-person
therapy, but they address several important barriers to dissemination of effective
depression treatments. Vigorous telephone outreach allowed us to engage patients
who might not be reached by traditional in-person treatment. Telephone sessions
eliminated travel and waiting time and allowed more flexible scheduling. Greater
privacy of telephone contacts helped to circumvent stigma. The telephone format
allowed therapists to use detailed agendas and checklists during therapy sessions,
an important contribution to treatment quality and consistency. While these
programs were tested in urban and suburban settings, advantages may be greater
in rural settings where access to psychotherapists is more limited and the
stigma attached to visiting a mental health provider may be greater.
Interpretation of these results should consider several limitations.
One third of potential participants declined to complete the initial assessment.
We cannot determine which specific elements of the telephone psychotherapy
program account for its effectiveness. Including all telephone and in-person
contacts, the telephone psychotherapy group received approximately 3 times
as many follow-up contacts as the usual care group. We cannot separate specific
content of the program (behavioral activation and cognitive restructuring)
from the increased contact provided to the telephone psychotherapy patients.
While the intervention programs may have increased rates of medication use,
the quality of pharmacotherapy in both groups was often still inadequate.
Finally, neither program would be supported by current fee-for-service reimbursement
The number of in-person follow-up visits in all 3 groups was approximately
3 visits over 6 months. The low rates of specialty mental health use are not
surprising given our exclusion of patients seeking or intending to seek specialty
care at enrollment, but the rates of primary care follow-up are disturbingly
low, even if one includes visits without a recorded mental health diagnosis.
Unfortunately, follow-up care after an antidepressant prescription is poor
in most US health care systems, with fewer than 25% of patients making a minimum
of 3 visits over 3 months.5 These low follow-up
rates were one of the primary motivations for developing the telephone care
management and telephone psychotherapy programs tested here.
Benefits of the telephone care management program appeared smaller than
those in our earlier trial,14 even though this
program was somewhat more intensive (3 contacts vs 2, inclusion of mailed
self-management workbook). This may reflect improvements in usual care since
the previous study. In that time, the health plan encouraged closer follow-up
of depression treatment and provided additional resources to primary care
teams (computerized registries, telephone follow-up scripts, new patient education
materials). Differences in sampling and eligibility between the 2 studies,
however, preclude a direct comparison of the 2 usual care groups.
Rates of participation in the telephone psychotherapy intervention were
markedly higher than those for traditional in-person therapies,7,35 even
compared with clinical trial participants selected for treatment motivation.36,37 Furthermore, we observed these high
rates of participation after excluding those already seeking psychotherapy.
The telephone psychotherapy program included significant outreach and a focus
on enhancing motivation for treatment. This level of therapist activity (eg,
multiple outreach calls including calls during evenings and weekends, repeated
mailings to patients not responding to telephone contacts) differs significantly
from the traditional office-based therapist's role. Such a public health approach—rather
than a traditional clinical approach—may be necessary to actually provide
empirically supported psychotherapy to the majority of depressed patients
not now served.
Primary care depression guidelines1,2 typically
do not recommend combining antidepressant medication and psychotherapy except
for patients with severe or chronic depression.38,39 Previous
studies have typically compared combined treatment to the pharmacotherapy
as provided in clinical trials (motivated patients, expert clinicians, frequent
follow-up visits, and high adherence rates). We evaluate the benefit of brief
structured psychotherapy added to typical primary care treatment (variable
motivation, infrequent follow-up, frequent nonadherence). We find that vigorous
efforts to engage patients in structured psychotherapy yield significant clinical
benefit over pharmacotherapy alone for primary care patients with moderate
depression. Given a shift in treatment patterns away from psychotherapy and
toward pharmacotherapy alone,40 our findings
have important implications for expanding the role of structured psychotherapy
in the care of depression.
Efforts to improve management of depression in primary care must consider
resource limitations and pressures to control costs. While we estimate the
cost of providing telephone psychotherapy to be less than $50 per session,
these additional resources should be directed to those patients most likely
to benefit. Post hoc analyses suggest that benefit was confined to those with
at least moderate depressive symptoms persisting for 2 to 4 weeks after a
first antidepressant prescription—approximately 65% of those starting
Our findings demonstrate the feasibility, acceptability, and effectiveness
of a telephone-based program including medication monitoring, care coordination,
and structured, depression-specific psychotherapy. For primary care patients
beginning antidepressant treatment, brief structured psychotherapy via telephone
adds significantly to usual care pharmacotherapy. These findings suggest the
need for a public health approach to psychotherapy emphasizing persistent
outreach and vigorous interventions to improve access to and motivation for