Efficacy of Nurse Telehealth Care and Peer Support in Augmenting Treatment of Depression in Primary Care


Background: Primary care treatment of depression needs
Objective: To evaluate the efficacy of 2 augmentations
to antidepressant drug treatment.
Design: Randomized trial comparing usual care, telehealth
care, and telehealth care plus peer support; assessments
were conducted at baseline, 6 weeks, and 6 months.
Setting: Two managed care adult primary care clinics.
Participants: A total of 302 patients starting antidepressant
drug therapy.
Interventions: For telehealth care: emotional support
and focused behavioral interventions in ten 6-minute calls
during 4 months by primary care nurses; and for peer support:
telephone and in-person supportive contacts by
trained health plan members recovered from depression.
Main Outcome Measures: For depression: the Hamilton
Depression Rating Scale and the Beck Depression Inventory;
and for mental and physical functioning: the
SF-12 Mental and Physical Composite Scales and treatment
Results: Nurse-based telehealth patients with or without
peer support more often experienced 50% improvement
on the Hamilton Depression Rating Scale at 6 weeks
(50% vs 37%; P=.01) and 6months (57% vs 38%; P=.003)
and on the Beck Depression Inventory at 6 months (48%
vs 37%; P=.05) and greater quantitative reduction in
symptom scores on the Hamilton scale at 6 months (10.38
vs 8.12; P=.006). Telehealth care improved mental functioning
at 6 weeks (47.07 vs 42.64; P=.004) and treatment
satisfaction at 6 weeks (4.41 vs 4.17; P=.004) and
6 months (4.20 vs 3.94; P=.001). Adding peer support
to telehealth care did not improve the primary outcomes.
Conclusion: Nurse telehealth care improves clinical outcomes
of antidepressant drug treatment and patient satisfaction
and fits well within busy primary care settings


DEPRESSION IS treated in

primary care almost as often

as in specialty care.1,2

Primary care treatment of

depression often has poor

quality3-12 and inadequate outcomes.3,4,6,12-15

Efforts to improve primary care treatment

of depression have included developing

and implementing clinical practice

guidelines15-22 and management strategies,

23-25 collaborative care models,8,25 team

care that incorporates the expertise of mental

health specialists,4,8,14 physician training

programs,8,17-26 therapy groups,27 telehealth

services,28-31 and the use of physician

extenders (Kathyrn Rost, PhD, et al, unpublished

data).4,29,32 No one model of care

has emerged as most effective.

In many primary care settings, particularly

in managed care, nurses have

played an important role in improving

care for chronic diseases, including

arthritis9 and hypertension.33 In a shortterm

inpatient psychiatric unit, nurse

telephone follow-up helped reduce readmissions.

34 Also, peer support has

been effectively used with patients with

cancer, schizophrenia, and substance



Patients were referred by physicians at the Hayward and San
Francisco primary care clinics within Kaiser Permanente
Northern California. Wall charts were placed in examining
rooms showing the Diagnostic and Statistical Manual of Mental
Disorders, Fourth Edition,41 criteria for major depressive
disorder; characteristics of patients who should be referred
directly to specialty care; and recommended starting doses
for selective serotonin reuptake inhibitor medications. Ninety
primary care physicians and 10 nurse practitioners received
2 hours of structured training and at least 1 hour of booster
training on the detection and management of depression.
Patients eligible for the study were diagnosed by a primary
care physician as having major depressive disorder
or dysthymia and given a prescription for a selective serotonin
reuptake inhibitor antidepressant. Study participants
were prescribed either fluoxetine hydrochloride or
paroxetine, reflecting prescribing practices in primary care
at the time of the study. Potential participants were excluded
if they received a previous antidepressant drug prescription
within the past 6 months, had an inadequate command
of the English language, reported current problems
with substance abuse, showed current suicide risk, or reported
thoughts of violence.
Patients were recruited during an 18-month period. Those
who were eligible and consented were randomly assigned
to 1 of 3 conditions: (1) usual physician care (physician counseling
and treatment with a selective serotonin reuptake inhibitor
medication), (2) usual physician care plus nurse telehealth
care, or (3) usual physician care plus nurse telehealth
care plus peer support. During the initial 9 months of recruitment,
patients were randomized only to conditions 1
or 2, with a 40% probability of assignment to condition 1.
During the remaining 9 months of recruitment, patients were
assigned to all 3 conditions, with 40% assigned to condition
1, 20% to condition 2, and 40% to condition 3. By design,
then, the overall study population was allocated 40%
to condition 1, 40% to condition 2, and 20% to condition 3;
thus, 40% of patients were assigned to an intervention without
telehealth care and 60% were assigned to an intervention
that included telehealth care. The allocation ratios were
designed to change over time so that we could begin the study
before we were prepared to deliver the peer intervention. Randomization
was stratified by facility. Consent and other procedures
of the study were approved by the institutional review
board of Kaiser Permanente.
Almost all of the patients referred to the study were recruited
through an in-person interview with a research assistant immediately
after the clinic visit that led to the referral. The rest
of the patients were recruited by telephone within 3 days of
referral. During the interview each patient’s eligibility was
checked, the study was explained, informed consent was obtained,
and baseline data were collected. Baseline measures
included the self-report version of the Hamilton Depression
Rating Scale,42-44 the Beck Depression Inventory,45 andthe SF-
measure perceived impairment of functioning due to mental
and physical disorders, respectively. Follow-up measures
were the same with 2 exceptions. The Hamilton Depression
Rating Scale–Interview49 was used rather than the self-report
version. Previous investigators44,49 found a high correlation
between the self-report and interview versions of the Hamilton
scale. The second exception was the deletion of the SF-
12PhysicalCompositeScale. Thismeasurewasincluded only
at baseline to control for global severity of physical illness.
Our primary measure of symptomatic outcome was the
Hamilton Depression Rating Scale. This instrument was expected
because it has shown this superiority in previous studies.50-53
Trained Kaiser Permanente interviewers (all graduate
students in psychology) assessed each patient by telephone
6 weeks and 6 months after study entry. The Beck Depression
Inventory was mailed to participants in advance, and their
answers were gathered via the telephone. All patients were
sought for interview at both follow-up times regardless of their
current treatment status, residence location, or membership
status in the Kaiser Permanente Health Plan.
Four interviewers were trained by the project clinical
director (J.F.M.). Training consisted of mock interviews in
which they were taught how to rate the Hamilton Depression
Rating Scale, followed by practice interviews in which
their ratings were reviewed, critiqued, and discussed. Training
on the other study instruments consisted of several mock
interviews that were observed and critiqued. Interviewers
were allowed to evaluate study patients only after they had
successfully completed training. During the study, 2 pairs
of interviewers periodically audited each other’s Hamilton
interviews and scored them independently. Interrater correlations
were 0.95 (n=31) and 0.97 (n=35)

Medication adherence data were gatheredfromcomputerized
pharmacy records showing every prescription filled by
studypatients ataKaiserPermanentepharmacy.Wemeasured
total milligrams of selective serotonin reuptake inhibitors dispensed
within 6 weeks and 6 months of randomization. We
excluded 15 patients at 6 weeks and 20 at 6 monthswhohad
terminated their Kaiser Permanente Health Planmembership
orhadfilled antidepressant prescriptions in a non–Kaiser Permanente
pharmacy up to the time of the analysis.
Usual Physician Care
Patients assigned to this condition continued to be seen as
needed by their primary care physician and, except for the
2 experimental interventions, could be referred for other
care as indicated. As previously described, most study physicians
had received at least 3 hours of training on the identification
and treatment of depression. They were asked to
continue the same pattern of follow-up visits with the patient
regardless of the study arm to which the patient was
assigned. It was our goal to have usual physician care represent
good care. We made it explicit that we did not intend
any reduced contact with the treating physician to “offset”
the cost of the supplemental interventions.
Nurse Telehealth Care
Nurse telehealth care consisted of 1 to 2 telephone calls per
week during the first 2 weeks of enrollment, 1 call per week
during weeks 3 to 8, and then 1 call every 2 weeks up to
week 16. The goal was to have 12 to 14 calls to each patient
during 16 weeks. Calls were limited to 10 minutes and
were scheduled in advance. Extra calls were permitted in
urgent circumstances, and the patient could leave telephone
messages for the nurse.
In each telephone call the nurse inquired about questions
the patient might have had about the antidepressant
medication, offered suggestions about how to deal with minor
side effects, and emphasized the importance of taking
the medication regularly. The nurse offered emotional support
and helped patients identify activities that they were
willing to try to be more active and to find pleasure. During
each telephone call the nurse reviewed the activities of
the previous week. With the patient, she developed a plan
for the next steps in doing these activities. Follow-up nurses
were members of the patient’s primary care clinic. This was
an advantage for patients because the nurse could also address
issues about other medical conditions and discuss the
patient’s overall health as well as his or her mental health.
By using regular clinic nurses we therefore hoped to improve
the overall care of the patient, not just the care for
depression. A log for each telephone call included current
medication and dosage, side effect problems, and a behavioral
plan. The nurse gave regular feedback on the progress
of each patient to the patient’s primary care physician.
Telehealth care nurses were chosen by a nursing supervisor
at each site based on the nurse’s interest in the
project and administrative staffing considerations. Fifteen
nurses completed a manualized 6-hour training workshop
developed and presented by the project clinical director
(J.F.M.). In treating study patients we used a single
trained nurse at each site with an alternate to cover leave
time. Nurses received ongoing weekly supervision from the
clinical director, a clinical psychologist, mostly by telephone
but with 1 visit to a site each month.
Peer Support
Peer support was provided by health plan members who
had experienced a successfully treated episode of major
depression or dysthymia and who volunteered to be
trained as peer supporters. In each case the volunteer’s
mental health provider endorsed his or her participation
in the program. Volunteers were recruited through
newspaper advertisements, notices posted in Kaiser Permanente
waiting rooms, and local television and radio
news coverage. All volunteers were carefully screened by
an experienced psychiatric social worker, who also
trained the volunteers, matched them to study patients,
and monitored their performance.
Training lasted approximately 20 hours and consisted
of lectures, role play, and discussion. Peer supporters
were supposed to model and share their successful
coping skills, provide emotional support, and encourage
self-monitoring and a continued connection to Kaiser Permanente
depression care. They were also expected to help
patients develop and keep a hopeful outlook.
Peers were linked with study patients of similar age
and sex who had been assigned to the peer support condition.
When possible, peers were also linked by similar
life experience such as job loss, divorce, or a similar medical
problem. Peers were expected to make a telephone or
in-person contact with the patient on 1 or more occasions,
continuing to contact the person to express an
interest in how they were doing for at least 6 months
after their assigned person entered the study. Incentives
such as movie tickets were offered to encourage face-toface
Dependent variables were total score on the Hamilton Depression
Rating Scale, total score on the Beck Depression
Inventory, 2 satisfaction factor scores, and the SF-12 Mental
Functioning Scale score. Consistent with previous research8,18,50-
52 on the treatment of depression in primary care,
results for the Hamilton Depression Rating Scale and the 

as 50% improvement rates. The rate was calculated by
assigning a score of 1 if the patient showed an improvement
of 50% or more from baseline and 0 if there
was less than a 50% improvement from baseline.
The analytic strategy was to focus sequentially
on the 3 study hypotheses. To test the first study hypothesis,
that nurse telehealth care leads to better outcomes
than usual physician care, we compared these
2 intervention groups on each outcome variable while
controlling for the baseline level of that outcome.
When comparing the means on quantitative variables,
we carried out an analysis of covariance with
usual physician care vs nurse telehealth care as the
independent variable and the baseline value of the
dependent variable as a covariate. We plotted residuals
and found that variances were homogeneous over
the range of predicted values. When analyzing the
50% improvement rates, we used logistic regression.
Again, we used the baseline value of the dependent
variable as a covariate. The 2 satisfaction factor
scores were evaluated using the Wilcoxon rank sum
test. We conducted all of these analyses separately
at 6 weeks and at 6 months.
There is a potential threat to validity from the
change in the randomization proportions made midway
through patient recruitment. To check for possible
bias, the analyses were re-run stratified by randomizationmethod.
method by treatment group interactions.
The second hypothesis, regarding the effect of
nurse telehealth care on medication adherence, was
tested using 3 interview questions covering current
antidepressant drug use. Computerized pharmacy data
were used to compute mean milligrams dispensed
from baseline to 6 weeks and from baseline to 6
months. Comparisons were made using 2 and t tests
as appropriate at each time.
The third hypothesis, regarding the effects on outcomes
of adding peer support to nurse telehealth care,
was tested by adding an additional contrast into the
previous analyses. We compared patients who received
nurse telehealth care plus peer support with
those who received nurse telehealth care alone.



abuse.35-40 We hypothesized that having nurses and
peers assist in the treatment of depression is an efficient
way to improve outcomes.
Wedeveloped a model for treating depression in primary
care that includes physician education and telephone
follow-up and support by trained primary care
nurses (“telehealth care”). In addition, we developed a
model of peer support provided by successfully treated,
formerly depressed health plan members. We aimed to
demonstrate high feasibility, easy implementation, and
improved patient outcomes and satisfaction.
We conducted a randomized trial comparing 3
models of care: usual physician care, including use of
selective serotonin reuptake inhibitor medication and
physician counseling; usual care plus telehealth care
provided by trained primary care nurses; and usual
care plus telehealth care plus peer support. These
treatments were implemented in 2 large adult primary
care clinics in the northern California area of the Kaiser
Permanente Medical Care Program, a nonprofit
group-model health maintenance organization. The
additive design enabled us to test 3 hypotheses: (1)
depressed patients who receive nurse telehealth care
experience greater reduction in depressive symptoms,
greater improvement in functioning, and greater satisfaction
with their care for depression compared with
those receiving usual physician care only; (2) favorable
outcomes are mediated by improved medication
adherence; and (3) the addition of peer support to
nurse telehealth care further improves these outcomes.
Four hundred eighty-six patients were referred to the
study; 116 were ineligible, 68 refused to give informed
consent, and 302 were enrolled. The main reasons for
ineligibility (some patients had several) included refusing
the recruitment interview (n=46), requiring referral
to other treatment (n=30), currently receiving an antidepressant
drug or psychotherapy (n=21), refusing the
prescribed selective serotonin reuptake inhibitor (n=8),
and not being literate in English (n=8). There was no
difference between the 184 nonparticipants and the 302
participants in age and sex, the only data available for
Women comprised 69% of the sample. The average
age was 55.4 years (range, 19-90 years). Patients were
white (63%), Hispanic (16%), African American (9%),
Asian (7%), and of other racial and ethnic backgrounds
(5%). The population was well educated: 90% were high
school graduates and 27% were college graduates. Household
income was reported to be less than $25000 by 34%
of the sample and greater than $60000 by 18%. More than
half of the participants lived with a spouse or partner,
and 22% lived alone. Full- or part-time employment was
reported by 53% of the sample, of whom 49% reported
that they held managerial or professional or technical
Using the unbalanced randomization described in
the “Treatment Assignment” subsection of the “Patients
and Methods” section, 123 patients (41%) were assigned
to usual physician care, 117 (39%) to nurse tele-

health care, and 62 (21%) to nurse telehealth care plus
peer support. Table 1 shows that there were no significant
differences between those assigned to nurse telehealth
care vs usual physician care on any of the demographic
variables or in baseline values of dependent
variables. Although peer support vs no peer support comparisons
within the telehealth care group are not emphasized
in this article, those subgroups were also similar
at baseline on all variables. The proportion of patients
initially prescribed fluoxetine vs paroxetine was almost
identical across all treatment groups.
Of 302 patients enrolled, we interviewed 271 (90%) at 6
weeks and 256 (85%) at 6 months. At 6 weeks, attrition
was due to 10 patients refusing to participate and interviewers
being unable to contact 21. At 6 months, attrition
was due to 13 patients refusing to participate, 5 having
physical illnesses that prevented participation, and
28 being unable to be contacted by interviewers. Logistic
regression analysis showed that attrition was not related
to treatment assignment at 6 weeks (Wald 2
P=.23) or 6 months (2
2=0.01; P=.90).
Nurse Telehealth Care
Eight of 179 patients randomized to nurse telephone
follow-up were unable to be reached by trained primary
care nurses. The nurses made a mean (SD) of 10.1 (3.6)
calls per patient to the 171 patients who spoke to a nurse
at least once during the 4 months of intervention. These
calls lasted a mean (SD) of 5.6 (2.3) minutes. The
estimated time for each telephone call, including callback
attempts and documentation, was 20 minutes.
We estimate that 20 patients could be contacted in a full
Peer Care
Of 62 patients randomized to peer support, 11 refused a
peer assignment and 9 never had a peer contact for other
reasons. Among the remaining 42 patients, 11 had 1 contact,
13 had 2 contacts, 14 had 3 to 5 contacts, and 4 had
9 to 20 contacts. Most patients were only contacted by
telephone, but 6 had at least 1 face-to-face contact.
Nurse Telehealth Care vs Usual Physician Care
Nurse telehealth care was hypothesized to be superior
to usual physician care with respect to reduced symptoms,
improved functioning, and greater satisfaction with
care for depression. These hypotheses were confirmed.
Table 2 shows the effect of nurse telehealth care on each
of these patient outcomes and the proportion of patients
who met the criterion of 50% improvement from
baseline on the Hamilton scale or the Beck Depression
Inventory. Fifty percent of patients receiving nurse telehealth
care experienced a 50% reduction in the Hamilton
scale score at 6 weeks compared with 37% receiving
usual physician care (P=.01). At 6 months, 57% of those
receiving nurse telehealth care showed a 50% improvement
in the Hamilton scale score compared with 38% in
usual physician care (P=.003). Hamilton Depression Rating
Scale quantitative scores showed a nonsignificant trend
favoring nurse telehealth care at 6 weeks and a significant
effect at 6 months (usual physician care=10.38, nurse
telehealth care=8.12; P.006 ). The 50% improvement
score for the Beck Depression Inventory shows an effect
at 6 months (usual physician care=37%, nurse telehealth
care=48%; P=.05).
Satisfaction with care showed a robust incremental
effect for nurse telehealth care. General satisfaction
with health care (factor 1) is significantly greater
among patients receiving nurse telehealth care at 6
weeks (nurse telehealth care=4.41, usual 

care=4.17; P=.004) and 6 months (nurse telehealth
care=4.20, usual physician care=3.94; P=.001). Satisfaction
with information provided about antidepressant
drugs (factor 2) shows a significant effect for
nurse telehealth care at 6 months (nurse telehealth
care=4.31, usual physician care=3.96; P.01).
The SF-12 Mental Functioning Scale reflecting “riskadjusted”
functioning48 impaired by mental symptoms
significantly favors nurse telehealth care at 6 weeks (nurse
telehealth care=47.07, usual physician care=42.64;
P=.004) but only a trend at 6 months (nurse telehealth
care=47.26, usual physician care=44.61; P=.10). Thus,
functioning improved earlier for the nurse telehealth care
group but both groups improved, and the usual physician
care group tended to catch up by 6 months.
Our midstream change in the randomization allocation
introduces a risk that differences between the
first and second cohort might bias the evaluation of
treatment effectiveness. To check whether there was
bias, we examined whether randomization cohort
interacted significantly with treatment assignment for
any outcome variable. Of the 7 outcome variables,
only the SF-12 Mental Functioning Scale showed an
interaction. The interaction was significant at 6 weeks
(P=.03) and a trend at 6 months (P=.10). In the early
cohort there was little advantage for nurse telehealth
care at 6 weeks but a large advantage at 6 months. In
the late cohort there was a large advantage for telehealth
care at 6 weeks, but the 2 treatment groups
converged by 6 months. The results shown in Table 2
are the average of these effects, which was significant
at 6 weeks and a trend at 6 months, both favoring
nurse telehealth care. Thus, the interaction with randomization
cohort does not change the SF-12 Mental
Functioning Scale findings.
Effect of Nurse Telehealth Care
on Medication Adherence
Contrary to our second hypothesis, we did not find that
medication adherence improved with nurse telehealth
care. At 6 weeks, 73% of the usual care patients and 80%
of those receiving nurse telehealth care reported taking
an antidepressant medication (P=.17), whereas at 6
months 54% in usual care and 56% in nurse telehealth
care were taking an antidepressant medication (P=.74).
For those still taking an antidepressant, there were no
significant differences in days missed in the past week
or days on which they took less than the prescribed dose.
Antidepressant dispensing data from our pharmacy database
show that treatment groups did not differ in mean
milligrams dispensed during the first 6 weeks of enrollment
(usual physician care=957, nurse telehealth care
plus peer support=867, t286=1.13; P=.26) or from enrollment
to 6 months (usual physician care=2267, nurse
telehealth care plus peer support=2111, t281=0.73; P=.45).
If anything, the trends suggest that patients receiving usual
care used more medication. When we repeated these
analyses in subgroups who were above or below a baseline
Hamilton Depression Rating Scale score of 19.75, we
also found no significant differences.
Added Value of Peer Support
Beyond Nurse Telehealth Care
The same variables shown in Table 2 were used to compare
outcomes at 6 weeks and 6 months between nurse
telehealth care and the combination of nurse telehealth
care and peer support. Of the 14 tests in Table 2, none
showed significant additive effects of peer support. There
were other outcome variables gathered that we thought 

were a few that showed effects at 6 months. These will
be presented elsewhere.
We sought an effective model for improving depression
treatment outcomes that could be implemented easily
within busy primary care settings. Our model uses the
nurses already in the primary care setting, builds on the
existing bond between primary care providers and patients,
and provides expert consultation and treatment
by mental health specialists when necessary. The model
uses medication, behavioral activation, education, brief
counseling, emotional support, monitoring of suicide risk,
and integration of depression care with ongoing care for
other chronic illnesses.
Our findings regarding nurse telehealth care have been
received with great interest by clinical leaders in the Kaiser
Permanente Health Plan and in other health maintenance
organizations. An implementation kit is being distributed
to other Kaiser Permanente regions as part of a
new depression disease management program developed
by Kaiser Permanente’s Care Management Institute.
We expected that one reason nurse telehealth care
would be superior to usual patient care is that it would increase
patients’ adherence to their prescribed medication.
We did not find this. Although further work is needed to
better understand the reasons for the effect, themechanism
seems to be more psychosocial than pharmacological.
Nurse telehealth care is significantly more effective
than usual care. But is the difference clinically important?
The most recent meta-analyses54,55 of antidepressant
drug treatment estimated a response rate of 50% for
antidepressant medications vs 32% for placebo in major
depressive disorder. Similarly, antidepressant treatment
for dysthymia produced a 59% response vs 37% for placebo.
Nurse telehealth care produces a 57% response rate
vs 38% for usual care. The difference between nurse telehealth
care and usual care is almost as large as that between
drug and placebo.
Comparisons specific to primary care can be made
using 2 studies by Katon et al.8,53 In one study (n=217),
usual care was compared with collaborative care that involved
increased physician visits during the first 4 to 6
weeks of treatment, including 2 visits by the primary care
physician and 2 or more by a psychiatrist. Eighty-nine percent
of all patients in this study were treated with tricyclic
antidepressant drugs. In the other study (n=153), usual
care was compared with a structured depression treatment
program provided in four to six 30-minute sessions
with doctoral-level psychologists that included behavioral
treatment to increase the use of adaptive coping strategies
and counseling to improve medication adherence.
The percentage of patients who started taking tricyclic antidepressant
drugs in this study was not reported but probably
included a larger proportion who had begun taking
selective serotonin reuptake inhibitors. In both studies, patients
were stratified by severity into those with major vs
minor depression, and the enhanced intervention showed
a clear advantage for major depression. It showed a much
smaller or nonsignificant advantage for minor depression.
Although it is difficult to compare the effects on depressive
symptoms that Katon et al8,53 obtained with those
of nurse telehealth care because different measures were
used, the size of their effects on major depression seems
similar to what we saw in our total sample of patients with
major depression or dysthymia.
Another difference was the intervention comparison.
Although all 3 studies differed in the specific content
of the interventions, both of the Katon et al8,53 interventions
involved considerable additional patient
contact by physicians or doctoral-level psychologists,
whereas the intervention reported herein used less staff
time, mostly of primary care nurses.
Peer support was operationally feasible in our setting
in the sense that we could recruit, train, and match a
sufficient number of qualified peers. However, unlike nurse
telehealth care, we did not set specific expectations for the
number and type of contacts between peers and patients.
As a result, only half of the patients randomized to peer
support had more than 1 contact with a peer and less than
10% had a face-to-face contact. This was less than we had
hoped for. Although we did not find that peer support improved
our primary outcomes in the presence of nurse telehealth
care, we think it might be worthwhile to explore
its value when it is more clearly structured and is the only
augmentation to usual physician care.
This is only the first evidence of the effectiveness
of nurse telehealth care, and the findings need to be replicated
in other circumstances to judge how confidently
we can recommend broad adoption. If nurse telehealth
care is broadly effective, its main advantage will be its
ease of implementation in primary care settings because
it does not require major staffing changes. We also believe
that nurse telehealth care can be improved with further
Accepted for publication April 18, 2000.
From the Division of Research, Kaiser Permanente
Northern California, Oakland (Mss Hunkeler, Groebe,
Braden, and Peng, Dr Hargreaves, and Mr Fireman); Departments
of Psychiatry (Dr Meresman) and Medicine (Dr
Getzell), Kaiser Permanente Northern California, Hayward;
Department of Psychiatry, University of California,
San Francisco (Dr Hargreaves); Department of Psychology,
Fordham University, Bronx, NY (Dr Berman); Smith-
Kline Beecham Pharmaceuticals (Dr Kirsch) and Department
of Psychiatry, Weill Medical College of Cornell
University, Westchester Division, White Plains, NY (Dr

Hurt); and Department of Medicine, Kaiser Permanente
Northern California, San Francisco (Dr Feigenbaum).
This research was supported in part by grants from the
Innovations Program of Kaiser Permanente Northern California,
Oakland, and the Community Services Program of
The Kaiser Permanente Medical Care Program–California
Division, and by an unrestricted education grant from Smith-
Kline Beecham Pharmaceuticals. This study is part of Kaiser
Permanente’s Depression Initiative, sponsored by the
Garfield Memorial Fund.
We thank the following individuals for their contributions
to this study: Elvira Z. Mann, RN, and Ann Green, RN,
of the Kaiser Permanente (KP) Medical Centers at Hayward
and San Francisco, Calif, respectively, for their provision
of telehealth care to the intervention group; A. Gerson
Schreiber, MD, and Raymond Zablotny, MD, of the KP
Medical Centers at Hayward and San Francisco, respectively,
for developing and leading the physician and nurse
practitioner training; Neil S. Handleman, MD, and Leslie
M. Lopato, MD, of the KP Medical Centers at Hayward and
San Francisco, respectively, for assisting with the physician
and nurse practitioner training; Viki Maxwell of the
KP National Member Technology Group for project coordination;
Luz de la Riva of the KP Division of Research in
Oakland, Calif, for successfully recruiting patients into the
study; Chanda Rankin and Alexandra Chase, MA, of the KP
Division of Research for successfully conducting follow-up
interviews with patients; David Cherry of the KP Division
of Research for data management and editorial assistance;
Stewart Proctor, PhD, of The Permanente Medical Group
Inc. Regional Operations Support Services at Oakland for
coordination with the Department of Psychiatry at Hayward;
Robin A. Dea, MD, chair of the Chiefs of Psychiatry,
KP Northern California, for cooperation and advice on clinical
and implementation matters throughout the course of the
study; Scott A. Bull, PharmD, of the KP Division of Research
for assistance with analysis and interpretation of medication
adherence data; Carol P. Somkin, PhD, of the KP Division
of Research for assistance in designing the peer support
intervention; A. Thomas McLellan, PhD, of the University
of Pennsylvania, Philadelphia, for thoughtful comments on
the study design; and Joe V. Selby, MD, of the KP Division
of Research for insightful editorial comments.
Reprints: Enid M. Hunkeler, MA, Division of Research,
Kaiser Permanente Northern California, 3505 Broadway,
Seventh Floor, Oakland, CA 94611-5463 (e-mail:



Case Comments:

  • No comments yet