CAMS.NDA.final

Full text

Research Article
DEPRESSION AND ANXIETY 28 : 963–972 (2011)
COLLABORATIVE ASSESSMENT AND MANAGEMENT OF
SUICIDALITY (CAMS): FEASIBILITY TRIAL FOR NEXT-DAY
APPOINTMENT SERVICES
Katherine Anne Comtois, Ph.D. M.P.H.,
1

David A. Jobes, Ph.D.,
2
Stephen S. O’Connor, Ph.D.,
1
David C. Atkins, Ph.D.,
1
Karin Janis, B.A.,
1
Chloe E. Chessen, B.A.,
1
Sara J. Landes, Ph.D.,
1,3
Anna Holen, M.D.,
1
and Christine Yuodelis-Flores, M.D.
1
Background: Despite the ubiquity of suicidality in behavioral health settings,
empirically supported interventions for suicidality are surprisingly rare. Given the
importance of resolving suicidality and therapists’ anxieties about treating suicidal
patients, there is a clear need for innovative services and clinical approaches.
The purpose of the current study was an attempt to address some of these needs by
examining the feasibility and use of a new intervention called the ‘ ‘Collaborative
Assessment and Management of Suicidality’’ (CAMS) within a ‘‘Next-Day
Appointment’’ (NDA) outpatient treatment setting.
Methods: As part of a larger
feasibility study , n 5 32 suicidal patients were randomly assigned to CAMS care
versus Enhanced Care as Usual (E-CAU) in an outpatient crisis intervention setting
attached to a safety net hospital. Intent to treat suicidal patients were seen and
assessed before, during, and after treatment (with follow-up assessments conducted at
2, 4, 6, and 12 months).
Results: The feasibility of using CAMS in the NDA setting
was clear; both groups appeared to initially benefit from their respective treatments
in terms of decreased suicidal ideation and overall symptom distress. Although
patients rated both treatments favorably, the CAMS group had significantly higher
satisfaction and better treatment retention than E-CAU. At 12 months post-
treatment, CAMS patients showed significantly better and sustained reductions in
suicidal ideation, overall symptom distres s, and increased hope in comparison to E-
CAU patients.
Conclusions: CAMS was both feasible in this NDA setting and
effective in treating suicidal ideation, distress, and hopelessness (particularly at 12
months followup). Depression and Anxiety 28:963–972, 2011.
rr
2011 Wiley
Periodicals, Inc.
Key words: suicide; attempted suicide; psychotherapy; risk assessment; crisis
intervention; feasibility studies; clinical trial
INTRODUCTION
More than 33,000 suicides occurred in the United
States in 2006—91 suicides per day or one suicide
every 16 min.
[1,2]
Death by suicide is part of a much
larger problem; millions of Americans have suicidal
thoughts and hundreds of thousands make suicide
attempts each year.
[3]
In 2008, 2.3 million people made
Published online 21 September 2011 in Wiley Online Library
(wileyonlinelibrary.com).
DOI 10.1002/da.20895
Received for publication 20 May 2011; Revised 25 July 2011;
Accepted 26 July 2011
The authors disclose the following financial relationships within
the past 3 years:
Contract grant sponsor: American Foundation for
Suicide Prevention.

Correspondence to: Katherine Anne Comtois, Harborview
Medical Center Box 359911, 325 9th Avenue, Seattle, WA 98104.
E-mail: comtois@uw.edu
Study conducted at Harborview Medical Center, 325 9th Avenue,
Seattle, WA 98104.
1
University of Washington, Seattle, Washington
2
The Catholic University of America, Washington, District of
Columbia
3
National Center for PTSD, VA Palo Alto Health Care System,
Palo Alto, CA
rr
2011 Wiley Periodicals, Inc.

a suicidal plan, 1.1 million made an attempt,
and 506,000 were hospitalized following a suicid e
attempt.
[4]
The lifetime cost of self-inflicted injuries in
the United States in 2000 was estimated at $33 billion,
including $1 billion for medical tre atment and $32
billion in lost productivity.
[5]
Patients who are hospitalized following an attempt
often encounter short-term inpatient hospitalization
focused heavily on treatment fo r men tal disorders. Some
individuals have ongoing outpatient care, but many do
not. It can be very difficult to find outpatient referral
options for suicidal inpatients. To address this, patients
without outpatient treat ment are sometimes given a
‘‘next-day appointment’’ (NDA) fo r follow-up post-
hospitalization. Clinicians pro viding NDA services must
have th e clinical tools to engage the patient, manage
suicidal risk, assist individuals to understand what leads
to their suicidal ity, and ultimately help them re-engage
with life. Yet studies indicate that many providers are
inadequately trained to provide proper assessment and
management of suicidal patients.
[6–10]
Although the current liter ature reflects a dearth of
suicide-specific treatment studies,
[11–13]
recent trials
show su bstantially reduced risk of su icide attempt
following cognitive therapy (CT) and dialectical behavior
therapy (DBT), respectively.
[14,15]
However, both require
substantial training that is particularly problematic in
crisis intervention clinics where there is significant
clinician turnover, reducing the likelihood that these
interventions will be available for NDA settings.
Preliminary evidence suggests a novel clinical inter-
vention called the Collaborative Assessment and
Management of Suicidality (CAMS)
[16]
may match
the specific dema nds of NDA settings. Jobes et al.
[17]
conducted a nonrandomized quasi-experimental clin-
ical trial of CAMS with suicidal US Air Force
personnel. Patients working with CAMS clinicians in
an outpatient setting markedly reduced their suicidal
ideation in significantly fewer sessions than patients
with Enhanced Care as Usual (E-CAU) clinicians.
Moreover, in the 6 months after initial mental health
treatment, CAMS patients had significantly fewer
emergency department (ED) visits and primary care
medical appointments and spent fewer total minutes in
these settings than E-CAU patients. Three additional
pilot studies utilizing within-subjects design methodology
have shown rapid reductions of suicidal ideation on
repeated measures with suicidal patients.
[18–20]
These initial studies are promising, but none
included a randomized design, a significant limitation
in the evaluation. In the present study, we have
conducted treatment development work and a small
randomized feasibility trial of CAMS versus E-CAU in
an NDA community setting. Our overall goal was
Stage 2B research as termed by Rounsaville et al.
[21]
using a hybrid of efficacy and effectiveness designs that
maximize external as well as internal validity compar-
able to larger trials at the Center for Healthcare
Improvement for Addictions, Mental Illness, and
Medically Vulnerable Populations (CHAMMP) at
Harborview Medical Center.
[22–30]
Thus, the aims of
our study were to demonstrate: (a) patient acceptance
of CAMS as provided in the type of clinic where it is
needed, (b) the ability to recruit sufficient numbers of
an externally valid target population, (c) the feasibility
of the treatments as provided in standard clinic with
standard therapists, (d) clinically significant improve-
ment over the course of treatment and followup on at
least one outcome domain, and (e) whether the
experimental treatment was likely to meet the thresh-
old of clinical significance (i.e. the effect size that might
‘‘disturb’’ equipoise in the specific clinical context
[31]
)
in a larger trial.
[21]
We predicted that CAMS would be
as feasible as E-CAU with better treatment retent ion
and patients’ satisfaction. Although this was a small
pilot study with the primary goal of feasibility, we
expected that CAMS would result in significant
reductions in suicidal ideation and psychological
distress and increases in reasons for living and hope
during treatment and at followup compared to E-CAU.
METHOD
SETTING
Harborview Medical Center is a county-owned hospital focused on
underserved and unfunded populations. Harborview’s Mental Health
Services (HMHS) provide long-term community mental health
services as well as a Crisis Intervention Service, Psychotherapy
Service, Gero-Psychiatry Service, and Peer Support. Study treatment
conditions were provided in the Crisis Intervention Service to which
all Harborview NDAs are referred.
PARTICIPANTS
At the outset of the study, potential participants were recruited
from the psychiatric emergency and consultation liaison psychiatry
services and later expanded to include inpatient psychiatry. Partici-
pants were 18 years or older and were evaluated by their treatment
provider for a recent suicide attempt or imminent risk. If the clinical
team determined that
(a) patient did not have appropriate outpatient mental health
treatment available for an appointment in the next 2 weeks
(b) a NDA and weekly outpatient followup was an appropriate
disposition plan, and
(c) the patient was sufficiently stable to be discharged home for a
minimum of 24 hours prior to NDA appointment, a clinician
contacted the study team.
Other inclusion criteria evaluated by the study screener were
consent to participate in all study procedures, significant suicidal
ideation according to Scale for Suicidal Ideation-Current (scor-
eZ13), and contact information for at least two long-term friends,
family, or service providers to help locate the participant for follow-
up assessments. To maximize external validity, this study sought to
exclude only those for whom participation in a voluntary research
study was not possible or for whom study treatment was not relevant.
Specifically, exclusion criteria were
(a) significant psychosis, cognitive or other impairment, or language
barrier such that the participant could not understand
study procedures or could not provide informed consent in
964 Comtois et al.
Depression and Anxiety

English, and thus psychosocial therapeutic care in English is
contra-indicated and
(b) a court order to treatment (e.g. subsequent to drug crime,
domestic violence, child abuse, driving while intoxicated) such
that their participation in study procedures could not be
considered voluntary and they would not be free to drop out
without negative consequences.
For ethical reasons, treating clinicians provided the initial
description of the study to appropriate patients, and when the patient
agreed, the potential participant was screened by study staff at the
earliest possible time. The screener discussed the study and
conducted informed consent with interested patients. For those
participants who consented and were eligible, the screener immedi-
ately conducted the pretreatment assessment, ran a minimization
random assignment program to determine treatment condition,
completed an appointment card with the appropriate contact
information for the participant’s NDA appointment and paid the
participant $20 for completing the pretreatment assessment.
TREATMENT CONDITIONS
The Collaborative Assessment and Management of Suicidality
(CAMS) is an intervention developed by the second author that
modifies how clinicians engage, assess, and treat suicidality.
[16,32]
.
CAMS creates the opportunity for a suicidal patient to identify the
‘‘drivers’’ or causes that lead to suicidal ideation and the subsequent
reduction in suicidal ideation and behavior as coping strategy. The
Suicide Status Form (SSF) guides assessment, treatment planning,
on-going tracking of risk, and outcome/disposition of care. The SSF
fosters collaboration as clinician and patient sit side-by-side and
deconstruct the patient’s suicidality through quantitative and
qualitative assessments and consideration of empirically based risk
factors (e.g. suicidal planning, access to means, attempt history). Each
subsequent CAMS session begins with core assessment items of the
SSF that focus the patient and therapist on suicide-related factors and
current risk and ends with a treatment plan that always includes a
crisis response plan. CAMS sessions were provided weekly for
50–60 min but there are no prescribed session-by-session format or
treatment strategies as long as the sessions are both collaborative and
suicide-focused. The therapist employs his/her own approach to
address the problems driving the suicidality. The length of CAMS is
determined by the time that it takes for suicidality to resolve
(operationally defined as three consecutive sessions of essentially no
suicidality). Thus, CAMS lasts a minimum of four sessions and, based
on the previous studies, resolution of suicidality usually occurs within
12 sessions.
Six CAMS clinicians (three case managers, two psychologists, and
one psychiatry resident) were recruited and participated in a 1-day
didactic training by Dr. Jobes, the CAMS developer. During the
study, weekly group consultation was conducted via phone. All
CAMS sessions were videotaped and sent to the research team at
Catholic University to determine adherence to CAMS during the
study. CAMS clinician adherence was defined as four out of the six
consecutive sessions receiving an overall rating of ‘‘satisfactory.’’ Two
CAMS clinicians did not reach acceptable levels of adherence before
dropping out during the training period (both case managers). The
remaining four clinicians achieved full CAMS adherence after an
average of 4.75 sessions with their first patient—that is within
2 months of training (three therapists and one case manager with
some therapy training; average years since degree M 5 5.8 SD 5 6.2).
Once full adherence was achieved, CAMS clinicians remained
adherent during 90% of subsequent rated sessions. (It is worth
noting that three out of the four CAMS clinicians were adherent to
CAMS treatment during 100% of all sessions rated.)
ENHANCED CARE AS USUAL
The goal of this study was to determine whether the implementa-
tion of CAMS is an improvement over standard outpatient crisis
intervention with a strong focus on external as well as internal
validity. Accordingly, we sought to insure the comparison condition
reflected standard policies and procedures for brief, crisis oriented,
mental health care. Standard care for a NDA patient at Harborview
entails an intake with the psychiatrist or psychiatric nurse practitioner
followed by 1–11 visits with a case manager (average years since
degree M 5 27.5, SD 5 3.5; mean 4.5 visits based on clinic data) and
as needed medication management. Treatment ends in 1–3 months
when the ‘‘crisis is resolved’’ with referral for primary care followup
or, when there is an appropriate diagnosis and funding is available,
additional mental health or substance abuse treatment. At the same
time, to minimize variability that would reduce internal validity, care
as usual in this study was enhanced in two ways: equivalent clinician
time was funded by the study and clinicians in both conditions were
asked to schedule a minimum of four sessions (i.e. the minimum
number of sessions in CAMS).
There were several treatment protocols common to CAMS and
E-CAU. Both conditions were implemented in the HMHS offices
following standard HMHS procedures including standard case
management, referral options, psychiatric evaluation, and medication
management, and referral to emergency or inpatient services if needed.
MEASURES
The Client Satisfaction Questionnaire (CSQ), an 8-item ques-
tionnaire, was administered post-treatment to assess patients’
satisfaction with treatment. This measure has been used frequently
for evaluating standard community mental health care
[33–35]
and was
supplemented by an open-ended question asking what was most and
least helpful about the study treatment.
The Scale for Suicide Ideation-Current (SSI)
[36]
is a 19-item
interviewer-administered scale that measures suicidal ideation at the
worst point of the past 2 weeks. It is a valid and reliable measure of
suicidal ideation for use with psychiatric patients
[36,37]
and has shown
predictive validity in an outpatient clinic sample for completed
suicide in the subsequent 10 years.
The Outcome Questionnaire-45 (OQ-45),
[38]
a 45-item ques-
tionnaire, measures key areas of psychological distress (symptoms,
interpersonal problems, and social role functioning). It is a widely
accepted tool for identifying, tracking, and measuring behavioral
health treatment outcomes
[39]
and was our measure of psychological
distress. It possesses good psychometric properties with adult
psychiatric patients
[38,40,41]
and the internal consistency in this
sample was good (a 5 .83).
The Reasons for Living Scale (RFL)
[42–44]
is a self-report 48-item
questionnaire with good psychometric properties that was used to
assess protective factors or beliefs buffering against suicidal behavior.
The RFL has several subscales, but only the total score was used in
the present study (a 5 .94).
The Optimism and Hope Scale (OHS)
[45,46]
is a 14-item self-
report measure combining a measure of dispositional optimism (Life
Orientation Test—Revised) and trait hopefulness (the Hope scale),
both of which show high reliability and construct validity.
[45–47]
Internal consistency for the combined scale in this study was very
good (a 5 .89).
The Suicide Attempt and Self-Injury Count SASI-C
[48]
is a very
brief survey of past self-inflicted injuries, categorized into suicide
attempts and nonsuicidal acts. The SASI-C creates counts of self-
inflicted injuries by method, medical risk severity, and lethality.
Participants received the lifetime assessment at pretreatment and the
follow-up version at all assessments.
965Research Article: CAMS Feasibility Trial for NDA Services
Depression and Anxiety

The Treatment History Interview—Short Form (THI)
[49]
uses a
time-line follow-back method of assessment to capture treatment
history including the number of ED visits, psychiatric and medical
hospital days, and physician and clinic visits. For participants who
reported hospitalization in the past year, previous studies revealed
90% agreement between participant report and hospital records for
number of admissions per participant, r 5 .99, and 80% agreement
for number of days per participant, r 5 .99. There were no false
positives for participation in individual psychotherapy (verified by
calling psychotherapists). Analyses of false negatives were not
possible.
Therapist adherence to CAMS was assessed with the CAMS
Rating Scale (CRS) an 18-item observer rating scale that was
completed by two of the authors (D.A.J. and S.S.O.) and a graduate
student trained by them based on videotapes of CAMS sessions.
PROCEDURES
All follow-up assessments were conducted in person or over the
phone (based on the participant’s preference) with a study assessor
blinded to treatment condition. Instruments were not modified for
phone follow-up assessments. Of the 77 follow-up assessments
(including partial interviews), 52% were conducted over the phone
(n 5 40); 48% were conducted in person (n 5 37). All assessors were
trained by the first author in all assessment protocols, including risk
assessment.
[50,51]
Given the small sample size, random assignment to
treatment condition was conducted via a minimization algorithm
matching for gender, history of suicide attempt, pre-existing use of
psychotropic medications, and history of substance abuse.
[52]
Follow-up assessments were conducted at 2, 4, 6, and 12 months
after the pretreatment assessment. Reimbursement was $10 for each
of the 2- and 4-month assessments and $15 for the 6- and 12-month
assessments; for each assessment, participants could also receive an
additional $5 for contacting the study to schedule the assessment and
another $5 for showing the first time the assessment was scheduled.
A minority of participants in both conditions received additional
assessments at 1, 3, and 5 months; however, for the purposes of this
article, analyses focus on the above time points most participants
completed. Primary outcome variables (suicidal ideation and
behavior, OQ-45, RFLS, OHS, and emergency services) were
assessed by a licensed clinician blind to treatment condition. The
treatment information (outpatient services, client satisfaction, etc.),
which would break the blind were assessed by the Research
Coordinator. The University of Washington Institutional Review
Board and a Data Safety Monitoring Board (DSMB) for the study
approved all study protocols, consents, and procedures.
DATA ANALYSES
Several features of the data needed to be taken into account in the
analyses. Data were imbalanced owing to missing data, and the
change across time for several outcomes was highly nonlinear, with a
rapid drop from baseline to early assessments, more gradual change
through 6 months, and notable differences at the 12-month follow-up
assessment. These data characteristics could be handled by (general-
ized) linear mixed models (GLMM
[53]
). GLMM model heterogeneity
across participants using random effects terms and can incorporate
imbalanced data (though, with assumptions about missing data).
However, most software to fit GLMM uses a maximum likelihood
framework for estimation, which can be biased in small samples.
Because of the small sample size with the present data, Bayesian
GLMM using Markov chain Monte Carlo (MCMC) estimation was
used, which has been shown to be unbiased even in small samples
(Raudenbush and Bryk,
[54]
Chapter 13).
As the change across time was notably nonlinear and different
between treatment groups, the following GLMM was used for the
primary outcomes:
Outcome
ti
¼b
0
1b
1
ðE-CAUÞ1b
2
ð2 mos Þ1b
3
ð4 mosÞ1b
4
ð6 mosÞ
1b
5
ð12 mos Þ1b
6
ðE-CAU  2 mosÞ
1b
7
ðE-CAU  4 mosÞ1b
8
ðE-CAU  6 mosÞ
1b
9
ðE-CAU  12 mosÞ1u
0i
1e
ti
where t indexes time and i indexes individuals. Assessment point was
modeled using a series of indicator variables, such that each
assessment predictor compares the mean at that assessment to
baseline. In addition, E-CAU (CAMS, 0; E-CAU, 1) is an indicator
variable for treatment. CAMS was chosen as the reference group so
that the within-group change for CAMS is captured by the main-
effect coefficients for assessment period. The two error terms (u
0i
, e
ti
)
capture the variability of individual participants and residual error,
respectively.
Bayesian models require prior distributions for each of the
parameters in the model, and noninformative priors were used for
all parameters (i.e. inverse-Wishart priors for random effects and
multivariate Normal prior for fixed effects). MCMC is a simulation-
based estimation procedure, and 100,000 MCMC iterations were run
following a burn-in of 50,000, with every 50th iteration saved for
analysis. Traceplots and the Gelman-Rubin statistic was used to
ascertain that models had converged (see Gelman and Hill
[55]
or
Lynch
[56]
for introductions to Bayesian methods). Analyses adhered
to the intent to treat principle, and participants were analyzed
according to their randomization. All analyses were done in R
v2.12.1
[57]
and made extensive use of the MCMCglmm package for
Bayesian GLMM.
[58]
RESULTS
FEASIBILITY OUTCOMES
Figure 1 shows the CONSORT flow diagram. Out
of 44 patients approached about the study, 32 were
randomized to either CAMS (n 5 16) or E-CAU
(n 5 16). Two CAMS patients turned out to be severe
and complex requiring intensive intervention and were
removed from the protocol based on consultation with
the IRB and DSMB, and one E-CAU participant was
court-ordered to treatment and therefore removed
from the protocol. The final sample was thus 14 CAMS
and 15 E-CAU. In total, 12 out of 14 CAMS patients
(78.5%) completed study assessments versus 10 out of
15 E-CAU patients (60%). (One in CAMS and three in
E-CAU dropped out of study assessments. The remain-
der no longer had valid contact informat ion and/or failed
to respond to over 40 contact attempts each at different
times of day on different days of the week.)
Demographic characteristics of the sample are
summarized in Table 1. The sample was 62% women,
66% Caucasian, and 69% single. A third were not
working at all and half the sample earned less than
$15,000/year. A total of 16% were street homeless and
19% only had a temporary place to live. There were no
significant differences on background characteristics
nor primary outcomes by treatment condition as
shown in Table 2 or Figure 2. (Note that the treatment
groups are not significantly different on suicide
966 Comtois et al.
Depression and Anxiety

attempts/nonsuicidal self-injuries at the start of treat-
ment as summarized in Table 2; Fisher’s exact test,
OR 5 3.5, 95% CI for OR 5 0.64, 21.69).
Twelve of the final 14 CAMS patients completed
treatment (86%) as did 10 of the final E-CAU
participants (67%). Participants were seen in CAMS
for an average of 8.4 (SD 5 4.4) sessions over 13.5
weeks and in E-CAU for an average of 4.7 (SD 5 3.1)
sessions over an average of 9.5 weeks. Average patient
satisfaction was high for both treatm ents but signifi-
cantly higher for the CAMS condition: E-CAU 3.1 and
CAMS 3.7 (on scale of 1, low through and 4, high;
t(24) 5 2.76 P 5.01). The majority of participants
completed the assessments as follows: Month 2 (52%
completed, 0% refused, 48% incomplete because could
not locate participant or he/she could not complete the
interview), 4 (66% completed, 0% refused, 36%
incomplete), 6 (69% compl eted, 7% refused, 24%
incomplete), and 12 (69% completed, 14% refused,
17% incomplete).
OUTCOME DESCRIPTIVES
Descriptive data for primary outcomes are shown in
Figure 2 by treatment condition and assessment point.
The descriptive data show that individuals who
received CAMS rapidly decreased their suicidal idea-
tion, which remained low throughout treatment and at
the 12-month assessment. E-CAU patients also steadily
reduced their suicidal ideation, though not quite as
rapidly, and there appears to be some rebound post-
treatment. A similar general pattern holds for overall
mental health as measured by the OQ-45. CAMS
patients rapidly improved and retained gains through
the 12-month assessment, whereas E-CAU patients
improved, though not as rapidly and with some
rebound. RFL and OHS both demonstrated some
early improvement for CAMS patients that were
largely maintained over 12 months, whereas the
E-CAU patients showed relatively less improvement
and at 12 months were similar to their baseline levels.
Given the low base rate of self-inflicted injury
and health services, no statistical analyse s were performed
but results are summarized for descriptive purposes in
Table 2. There were only three suicide attempts postbase-
line—2 in CAMS and one in E-CAU. Including both
suicide attempts and nonsuicidal self-injuries in a self-
inflicted injury total, CAMS participants made slightly
fewer self-inflicted injuries at all points—including base-
line. CAMS also had slightly fewer ED admissions overall
and ED admissions for behavioral health reasons.
Inpatient psychiatric days varied over the time period
without clearly favoring one condition.
BAYESIAN GLMM ANALYSES
Table 3 lists the coefficients and 95% CI for the
Bayesian GLMM presented above for each of the four
primary outcomes. Note that SSI was strongly skewed
and bounded by zero. This measure refl ects a count of
19 characteristics of an individual’s plans and wishes to
die by suicide, and thus SSI was modeled as a Poisson
GLMM, whereas other outcomes were fit using linear
mixed models assuming normally distributed errors.
Figure 1. Consort diagram.
967Research Article: CAMS Feasibility Trial for NDA Services
Depression and Anxiety

The key difference for interpretation is that the
Poisson GLMM uses a log link function, and similar
to a logistic regression model, raw coefficients are
typically exponentiated (i.e. raised to base e)and
interpreted as rate ratios (RR). Confidence intervals
for RRs excluding 1 are significant at the Po.05 level.
The main effect of treatment in all models tests the
difference between treatments at baseline, which is
nonsignificant for all primary outcomes.
The results show that CAMS patients were signifi-
cantly improved from their baseline levels of SSI,
ranging from a 61% reduction at 2 months (RR 5 0.39)
to an 89% reduction at 12 months (RR 5 0.11) (as
noted above, E-CAU is coded 1 and hence the main
effects of assessment represent the within group change
of CAMS patients). The E-CAU by assessment
interactions at 2 and 4 months show that E-CAU
patients were reporting more suicidality than CAMS
patients, though not significantly so. At the 6-month
assessment, the two groups were virtually identical in
reported SSI, whereas at 12 months E-CAU patients
were notably (and significantly) worse than CAMS
patients (RR 5 4.81), reflecting what was seen in the
descriptive data.
Figure 2. Means and 80% confidence intervals for primary
outcomes by treatment condition and assessment point.
TABLE 1. Demographics of sample
N % M7SD Range
Age 36.8710.1 19–62
Sex Female 18 62
Ethnicity White-Caucasian 19 66
Black-African American 4 14
Asian or Asian American 1 3
Latino/Latina 1 3
Mixed or other ethnicity 3 10
Participant does not know 1 3
Marital status Single, never married 20 69
Divorced or separated 7 24
Widowed 1 3
Married 1 3
Sexual orientation Heterosexual 24 83
Bisexual 5 17
Homosexual 0 0
Residential arrangement Street homeless 5 16
Temporary housing 6 19
Permanent housing 21 65
Employment status Not working or retired 11 38
o20 h per week 3 10
20–39 h per week 5 17
Z40 h per week 10 35
Education Some high school 3 10
High school degree or GED 10 35
Some college 8 28
Business or technical training 3 10
College degree 5 17
Annual income o$5,000 8 28
$5,000–$14,999 8 28
$15,000–$29,999 8 28
>$30,000 4 14
Unknown 1 2
968 Comtois et al.
Depression and Anxiety

The results for the OQ-45 are quite similar to those
for SSI: CAMS patients rapidly improved, with
significant reductions vs. baseline at 4-, 6-, and
12-month assessments. The E-CAU by assessment
interactions were all positive (suggesting worse mental
health for E-CAU) and were significantly different at
12 months. The RFL scale was notable in that the
intraclass correlation coefficient (ICC) was extremely
high (.80). This reflects that scores within an individual
were very highly correlated; perhaps suggesting that
RFL is more trait-like as opposed to something that is
readily manipulable by psychological intervention.
Although the assessment coefficients were all positive
for CAMS, none of them were significantly above
baseline levels. Similarly, none of the E-CAU by
assessment contrasts was significant. The OHS scale
shows that the CAMS patients had significantly greater
hope beginning at the 4-month assessment and that
E-CAU patients were significantly below their CAMS
counterparts at 12 months.
DISCUSSION
The current feasibility trial using a hybrid efficac y
and effectiveness design found that patients psychia-
trically hospitalized for a suicide attempt or imminent
risk were interested in and willing to participate in
outpatient NDA treatment and CAMS. The treatments
TABLE 2. Descriptive statistics of infrequent outcomes by treatment condition and assessment
Outcome variable Assessment time point
CAMS E-CAU
Mean SD Mean SD
Suicide attempts/self-inflicted injuries 0 month 3.0 9.3 7.7 24.5
2 month N/A N/A 5.5 7.8
4 month 0.0 0.0 0.8 1.8
6 month 0.2 0.4 0.0 0.0
12 month 1.2 3.9 3.3 7.6
ED admissions (all) 0 month 1.5 1.2 1.6 0.8
2 month N/A N/A 0.5 0.7
4 month 0.4 0.5 0.4 0.7
6 month 0.4 0.5 0.2 0.4
12 month 0.4 0.8 1.0 2.4
Behavioral health ED admissions only 0 month 1.3 1.1 1.1 0.6
2 month N/A N/A 0.5 0.7
4 month 0.1 0.4 0.4 0.7
6 month 0.2 0.4 0.2 0.4
12 month 0.2 0.4 0.6 1.6
Number of inpatient days 0 month 5.5 5.4 7.0 7.0
2 month N/A N/A 4.0 5.7
4 month 1.4 2.5 1.0 2.3
6 month 3.5 7.0 1.3 4.6
12 month 1.4 4.5 3.2 8.0
N/A, not applicable (as no data observed for specific treatment at given time point).
TABLE 3. Regression results (coefficients and 95% confidence intervals) for primary outcomes
SSI OQ-45 RFL OHS
RR Lower Upper B Lower Upper B Lower Upper B Lower Upper
Intercept 24.40 22.14 26.89 85.61 74.67 100.94 169.52 144.82 195.97 33.84 29.53 37.52
E-CAU 0.94 0.79 1.11 4.08 10.36 25.59 7.16 39.14 29.47 2.55 8.01 2.75
2 mos 0.39

0.20 0.77 9.69 30.67 3.00 8.54 15.38 31.86 3.54 0.17 7.13
4 mos 0.27

0.13 0.54 20.63

32.44 7.65 12.11 6.53 30.96 4.96

1.97 7.81
6 mos 0.32

0.15 0.66 20.93

34.72 11.44 5.71 11.56 25.21 4.15

1.12 6.83
12 mos 0.11

0.04 0.30 32.27

45.50 20.49 8.04 9.17 28.62 5.60

2.61 8.44
E-CAU  2 mos 1.64 0.72 3.71 0.77 23.46 24.29 16.13 47.30 16.66 0.11 4.91 5.20
E-CAU  4 mos 1.84 0.82 4.14 2.41 14.65 18.60 5.65 28.35 21.40 2.83 6.72 1.25
E-CAU  6 mos 0.96 0.36 2.60 6.92 8.55 25.39 5.65 20.92 30.73 1.63 5.69 2.33
E-CAU  12 mos 4.81

1.61 14.33 19.65

4.13 36.65 2.78 30.89 21.75 4.07

8.77 0.12
E-CAU is an indicator variable (CAMS, 0; E-CAU, 1), assessment covariates are indicator variables comparing reported assessment to baseline
(e.g. 2 mos: 0, baseline assessment; 1–2 month assessment). SSI was modeled as a Poisson-generalized linear mixed model. RR, rate ratio, B,
unstandardized regression coefficient, lower and upper report bounds of 95% confidence interval.

Statistically significant difference.
969Research Article: CAMS Feasibility Trial for NDA Services
Depression and Anxiety

were found to be acceptable to participants, albeit with
a statistically higher satisfaction for CAMS vs. E-CAU.
We found that for clinicians with psychotherapy
training, we could train CAMS to adherence quickly
and that CAMS ther apists maintained adherence. Case
management training alone was insufficient training for
CAMS adherence.
We noted two other feasibility issues that sugge st
modifications for future trials. First, the majority of
suicidal individuals in this study had both high
symptoms and high service needs that frequently
required long-term follow-up case management, phar-
macotherapy, or other treatment. It appeared that the
NDA structure of stabilization with study treatments
then referral was a difficult transition for these patients.
A second related issue was the need for case manage-
ment for housing, finances, vocational, and medical
problems. These patients varied widely in their need
for these services, but often needs were beyond the
capacity of the CAMS session or took up all the time of
the E-CAU clinician. In response to both of these
issues, a model that can pair NDA patients with the
level of case management services they need and begins
long-term care (when needed) simultaneous with NDA
treatment would likely lead to a more effective
transition clinically—although clinics would have to
overcome structural hur dles in terms of access criteria
and funding for such programs.
Results showed that CAMS participants had im-
proved more than E-CAU at the 12-month assessment
on suicidal ideation, me ntal health symptoms, and
hope; with CAMS pa rticipants continuing improve-
ment to near zero where E-CAU participants appeared
to lose ground back into the clinical range. Patients in
the study clearly took to the CAMS approach; they
were both satisfied and showed good retention to the
treatment. The overt CAMS focus on reducing suicidal
ideation and behavior as a means of coping with pain
and distress and strengthening of alternative coping
skills clearly seemed to work for many of these patients.
In addition, the CAMS patients were responsive to the
idea of identifying, targeting, and treating the drivers of
their suicidality with collaboratively developed pro-
blem-focused interventions. Statistically significant
effects were seen even with modest sample sizes and
statistical power (particularly at 12-month followup),
speaking to the potentially meaningful and end uring
impact of the approach.
There are a number of limitations to this study. The
sample was small and recru ited almost entirely from
inpatient psychiatric services looking for follow-up care
for their patients without needed funding. Thus, this
population is not representative of all individuals who
make suicide attempts or are imminently suicidal and
larger trials are clearly needed. However, this was the
group of high concern to the inpatient psychiatric units
who were worried they would be readmitted to the ED,
and they referred many more patients than we could
accept and continued to try and refer patients to the
study long after the study had closed. Secon d, because
this study was a feasibility trial, several study proc e-
dures changed over the course of the trial. In particular,
the change from recruitment in emergency and
consultation-liaison psychiatry services which were
the initial goal of the study to inpatient psychiatry
services and the change in the assessment time points.
These changes emphasize the need for replication of
these results. There were also problems locating
participants for follow-up and facilitating their com-
pletion of assessments despite considerable efforts. We
could discern no pattern of who did not complete
assessments.
Third, patients in the CAMS condition received on
average a greater number of treatment sessions, which
may have affected the treatment outcomes above and
beyond the specific effect of the therapeutic interven-
tion. The study’s funding of equivalent clinical time for
both conditions and the fact that one E-CAU clinician
had longer and more frequent sessions than the other
suggests that this difference was clinical choice rather
than owing to practical considerations. Although this
potential confound reflects our study’s bias toward
external validity of both control group and subsequent
results, replication controlling for attention is required
to rule out this limitation. Finally, the CAMS inter-
ventionists were composed of three therapist s and one
case manager with some therapy training, whereas the
E-CAU interventionists were both case managers,
which may have affected the results. Agai n, this
potential confound reflects our emphasis on external
validity and future studies with comparable providers
are needed to rule it out. However, given the over-
whelming case management needs of many study
participants, the more experienced case management
skills of E-CAU therapists compared to CAMS
therapists may have offered important benefits. As
noted above, pairing CAMS with ongoing case
management services would most likely maximize the
likelihood of significant overall improvement in the
patients’ lives. The lead authors are currently conduct-
ing a large r well-powered randomized clinical trial
of CAMS; clearly, other such trials are needed.
In additio n, other strategies to improve access to
outpatient followup after admission for suicidal beha-
vior—either NDA or otherwise—is critical as are
improvements to the quality of this care.
CONCLUSION
After many years of research neglect, this is an
exciting time of expanding empirical studies related to
the treatment of suicidal risk. Recent innovations in
screening, more thorough assessments, the use of safety
plans, and evidence-base d suicide-specific interven-
tions are increasingly being seen in the professional
literature and in clinical practice. Based on the current
investigation—even with extremely disadvantaged and
challenged suicidal patients—the prospect of making a
970 Comtois et al.
Depression and Anxiety

meaningful and potentially enduring difference using
CAMS seems possible.
Acknowledgments. This study was funded by a
Distinguished Investigator award from the American
Foundation for Suicide Prevention and by in kind
resources provided by Harborview Mental Health
Services. The authors thank them both very much for
their support.
REFERENCES
1. National Center for Injury Prevention and Control C: Web-
based Injury Statistics Query and Reporting System (WIS-
QARS). Edited by Prevention CfDCa, Available at:
www.cdc.gov/injury/wisqars/index.html., 2007.
2. Kung H-S, Hoyert DL, Xu J, et al. Deaths: Final data for 2005.
National Vital Statistics Reports 56: Obtained from http://
www.cdc.gov/nchs/data/nvsr/nvsr56/nvsr56_10.pdf. January 16,
2008.
3. McIntosh J. USA suicide: 2003 Official final data. Fact sheet
developed by the American Association of Suicidology from
National Center Health Statistics www.suicidology.org, 2006.
4. Substance Abuse and Mental Health Services Administration,
Office of Applied Studies. (September 17, 2009). The NSDUH
Report: Suicidal Thoughts and Behaviors among Adults.
Rockville, MD.
5. Corso PS, Mercy JA, Simon TR, et al. Medical costs and
productivity losses due to interpersonal and self-directed violence
in the United States. Am J Prev Med 2007;32:474–482.
6. Fang F, Kemp J, Jawandha A, et al. Encountering patient suicide:
a resident’s experience. Acad Psychiatry 2007;31:340–344.
7. Office of Surgeon General: Office of the Surgeon General:
National strategy for suicide prevention: goals and objectives for
action. Edited by Services CfMH, U.S. Dept. of Health and
Human Services, Public Health Service, 2001.
8. Bongar B, Harmatz M. Clinical psychology graduate education
in the study of suicide: availability, resources, and importance.
Suicide Life Threat Behav 1991;21:231–244.
9. Jobes DA, Rudd MD, Overholser JC, et al. Ethical and
competent care of suicidal patients: contemporary challenges,
new developments, and considerations for clinical practice. Prof
Psychol Res Pract 2008;39:405–413.
10. Dewar I, Eagles J, Klein S, et al. Psychiatric trainees’ experiences
of, and reactions to, patient suicide. Psychiatr Bull 2000;24:20–23.
11. Linehan MM. Suicide intervention research: a field in desperate
need of development. Suicide Life Threat Behav 2008;38:
483–485.
12. Comtois KA, Linehan MM. Psychosocial treatments of suicidal
behaviors: a practice-friendly review. J Clin Psychol 2006;62:
161–170.
13. Tarrier N, Taylor K, Gooding P. Cognitive-behavioral interven-
tions to reduce suicide behavior: a systematic review and meta-
analysis. Behav Modif 2008;32:77–108.
14. Brown GK, Ten Have T, Henriques GR, et al. Cognitive therapy
for the prevention of suicide attempts: a randomized controlled
trial. J Am Med Assoc 2005;294:563–570.
15. Linehan MM, Comtois KA, Murray AM, et al. Two-year
randomized controlled trial and follow-up of dialectical behavior
therapy vs therapy by experts for suicidal behaviors and border-
line personality disorder. Arch Gen Psychiatry 2006;63:757–766.
16. Jobes DA. Managing Suicidal Risk: A Collaborative Approach.
New York: The Guilford Press; 2006.
17. Jobes DA, Wong SA, Kiernan A, et al. The collaborative
assessment and management of suicidality vs. treatment as usual:
a retrospective study with suicidal outpatients. Suicide Life
Threat Behav 2005;35:483–497.
18. Arkov K, Rosenbaum B, Christiansen L, et al. Treatment of
suicidal patients: the collaborative assessment and management
of suicidality. Ugeskr Laeger 2008;170:149–153.
19. Ellis T, Allen, JG, Lamb, D, Conaway, E, Daza, P, Herrera, S.
Outcomes from a Pilot Study of CAMS in a Psychiatric Inpatient
Setting in Poster presented at the 43rd American Association of
Suicidology Annual Conference. Orlando, FL, 2010.
20. Jobes DA, Kahn-Greene E, Greene JA, et al. Clinical improve-
ments of suicidal outpatients: examining suicide status form
responses as predictors and moderators. Arch Suicide Res
2009;13:147–159.
21. Rounsaville BJ, Carroll KM, Onken LS. A stage model of
behavioral therapies research: getting started and moving from
stage one. Clin Psychol Sci Pract 2001;8:133–142.
22. Krupski A, Sears JM, Joesch JM, et al. Impact of brief
interventions and brief treatment on admissions to chemical
dependency treatment. Drug Alcohol Depend 2010;110:
126–136.
23. Roy-Byrne P, Craske MG, Sullivan G, et al. Delivery of evidence-
based treatment for multiple anxiety disorders in primary care: a
randomized controlled trial. J Am Med Assoc 2010;303:
1921–1928.
24. Zatzick DF, Roy-Byrne P, Russo J, et al. A randomized
effectiveness trial of stepped collaborative care for acutely injured
trauma survivors. Arch Gen Psychiatry 2004;61:498–506.
25. Zatzick DF, Simon GE, Wagner AW. Developing and imple-
menting randomized effectiveness trials in general medical
settings. Clin Psychol Sci Pract 2006;13:53–68.
26. Zatzick D, Galea S. An epidemiologic approach to the
development of early trauma focused intervention. J Trauma
Stress 2007;20:401–412.
27. Roy-Byrne P, Craske M, Stein M, et al. A randomized
effectiveness trial of cognitive-behavioral therapy and medication
for primary care panic disorder. Arch Gen Psychiatry
2005;62:290–298.
28. Gentilello LM, Rivara FP, Donovan DM, et al. Alcohol
interventions in a trauma center as a means of reducing the risk
of injury recurrence. Ann Surg 1999;230:473–480; discussion
480–483.
29. Terrell F, Zatzick DF, Jurkovich GJ, et al. Nationwide survey of
alcohol screening and brief intervention practices at US level I
trauma centers. J Am Coll Surg 2008;207:630–638.
30. Ciechanowski P, Wagner E, Schmaling K, et al. Community-
integrated home-based depression treatment in older adults: a
randomized controlled trial. J Am Med Assoc 2004;291:
1569–1577.
31. Kraemer HC, Mintz J, Noda A, et al. Caution regarding the use
of pilot studies to guide power calculations for study proposals.
Arch Gen Psychiatry 2006;63:484–489.
32. Jobes DA, Comtois KA, Brenner LA, et al. Clinical trial
feasibility studies of the collaborative assessment and manage-
ment of suicidality (CAMS). In: O’Connor R, Platt S, Gordon J,
editors. International Handbook of Suicide Prevention: Research,
Policy & Practice. London: Wiley; 2011.
33. Attkisson CC, Zwick R. The Client Satisfaction Questionnaire:
psychometric properties and correlations with service utilization
and psychotherapy outcome. Eval Program Plann 1982;5:
233–237.
34. Gaston L, Sabourin S. Client satisfaction and social desirability
in psychotherapy. Eval Program Plann 1992;15:227–231.
971Research Article: CAMS Feasibility Trial for NDA Services
Depression and Anxiety

35. Nguyen T, Attkisson C, Stegner B. Assessment of patient
satisfaction: development and refinement of a service evaluation
questionnaire. Eval Program Plann 1983;6:299–314.
36. Beck AT, Brown GK, Steer RA. Psychometric characteristics of
the Scale for Suicide Ideation with psychiatric outpatients. Behav
Res Ther 1997;35:1039–1046.
37. Beck AT, Steer RA, Ranieri W. Scale for Suicide Ideation:
psychometric properties of a self-report version. J Clin Psychol
1988;44:499–505.
38. Lambert MJ, Hansen NB, Umphress V, et al. Administration and
scoring manual for the Outcome Questionnaire (OQ 45.2).
Wilmington, DE, American Professional Credentialing Services,
1996.
39. Maruish ME. Psychological testing in the age of managed
behavioral health care Mahwah, N.J., L. Erlbaum As sociates, 2001.
40. Lambert MJ, Burlingame G, Umphress V, et al. The reliability
and validity of the Outcome Questionnaire. Clin Psychol
Psychother 1996;3:249–258.
41. Umphress VJ, Lambert MJ, Smart DW, et al. Concurrent and
construct validity of the Outcome Questionnaire. J Psychoedu-
cational Assess 1997;15:40–55.
42. Linehan MM, Goodstein JL, Nielsen LS, et al . Reasons for staying
alive wh en you are thinking of killing yourself: the Reasons for
Living Inve ntory. J Consult Cl in Psychol 1983;51:276–286.
43. Dobrov E, Thorell LH. ‘‘Reasons For Living’’—translation,
psychometric evaluation and relationships to suicidal behaviour in
a Swedish random sample. Nord J Psychiatry 2004;58:277–285.
44. Osman A, Kopper BA, Linehan MM, et al. Validation of the adult
suicidal ideation questionnaire and the reasons for living
inventory in an adult psychiatric inpatient sample. Psychol
Assess 1999;11:115–123.
45. Scheier MF, Carver CS, Bridges MW. Distinguishing optimism
from neuroticism (and trait anxiety, self-mastery, and self-
esteem): a reevaluation of the Life Orientation Test. J Pers Soc
Psychol 1994;67:1063–1078.
46. Snyder CR, Harris C, Anderson JR, et al. The will and the ways:
development and validation of an individual-differences measure
of hope. J Pers Soc Psychol 1991;60:570–585.
47. Hirsch JK, Britton PC, Conner KR. Psychometric evaluation
of the life orientation test—revised in treated opiate
dependent individuals. Int J Ment Health Addict 2010;8:
423–431.
48. Linehan MM, Comtois KA: Lifetime Parasuicide History,
Available Department of Psychology. Seattle, WA: University
of Washington; 1996.
49. Linehan MM, Heard HL: Treatment History Interview (THI),
Available Department of Psychology. Seattle, WA: University of
Washington; 1987.
50. Linehan M, Comtois K, Ward-Ciesielski E. Assessing
and managing risk with suicidal individuals: the UW Risk
Assessment Protocol (UWRAP) and UW Risk Assessment and
Management Protocol (UWRAMP), University of Washington,
2010.
51. Reynolds SK, Lindenboim N, Comtois KA, et al. Risky
assessments: participant suicidality and distress associated with
research assessments in a treatment study of suicidal behavior.
Suicide Life Threat Behav 2006;36:19–34.
52. Scott NW, McPherson GC, Ramsay CR, et al. The method of
minimization for allocation in clinical trials: a review. Control
Clin Trials 2002;23:662–674.
53. Hedeker D, Gibbons R. Longitudinal Data Analysis. New York:
Wiley; 2006.
54. Raudenbush SW, Bryk AS: Hierarchical Linear Models: Applica-
tions and Data Analysis Methods. Thousand Oaks, CA: Sage
Publications; 2002.
55. Gelman A, Hill J. Data Analysis Using Regression and Multi-
level/Hierarchical Models. Cambridge, MA: Cambridge Uni-
versity Press; 2007.
56. Lynch SM. Introduction to Applied Bayesian Statistics and
Estimation for Social Scientists. New York: Springer; 2007.
57. R Development Core Team: R: A language and environment
for statistical computing. Vienna, Austria, R Foundation for
Statistical Computing. http://www.R-project.org, 2010.
58. Hadfield J. MCMC methods for multi-response generalized
linear mixed models: The MCMCglmm R package. J Stat Softw
2010;33:1–22.
972 Comtois et al.
Depression and Anxiety