286 JOURNAL OF STUDIES ON ALCOHOL AND DRUGS / MARCH 2011
A Randomized Controlled Trial of Cognitive-Behavioral
Treatment for Depression Versus Relaxation Training
for Alcohol-Dependent Individuals With Elevated
RICHARD A. BROWN, PH.D.,† SUSAN E. RAMSEY, PH.D.,† CHRISTOPHER W. KAHLER, PH.D.,† KATHLEEN M. PALM, PH.D.,†
PETER M. MONTI, PH.D.,† DAVID ABRAMS, PH.D.,† MARYELLA DUBREUIL, M.S.,† ALAN GORDON, M.D.,
AND IVAN W. MILLER, PH.D.
Department of Psychiatry and Human Behavior, Butler Hospital/Brown Medical School, 345 Blackstone Boulevard, Providence,
Rhode Island 02906
ABSTRACT. Objective: A previous pilot study found positive out-
comes among alcohol-dependent individuals with elevated depressive
symptoms who received cognitive-behavioral treatment for depression
(CBT-D; n = 19) compared with a relaxation training control (RTC;
n = 16). The current study represents a replication of this pilot study
using a larger sample size and a longer follow-up assessment period.
Method: Patients entering a partial hospital drug and alcohol treatment
program who met criteria for alcohol dependence and elevated depres-
sive symptoms (Beck Depression Inventory score ≥ 15) were recruited
and randomly assigned to receive eight individual sessions of CBT-D (n
= 81) or RTC (n = 84). Results: There were signifi cant improvements
in depressive and alcohol use outcomes over time for all participants.
Compared with RTC, the CBT-D condition had signifi cantly lower levels
of depressive symptoms, as measured by the Beck Depression Inventory,
at the 6-week follow-up. However, this effect was inconsistent because
there were no differences in the Modifi ed Hamilton Rating Scale for
Depression between conditions at that time point and there were no sig-
nifi cant differences at any other follow-up. No signifi cant between-group
differences on alcohol use outcomes were found. Conclusions: The
current fi ndings did not replicate the positive outcomes observed in the
CBT-D condition in our previous pilot study. Possible explanations for
why these fi ndings were not replicated are discussed, as are theoretical
and clinical implications of using CBT-D in alcohol treatment. (J. Stud.
Alcohol Drugs, 72, 286-296, 2011)
Alcoholism grant AA10958 awarded to Richard A. Brown.
or via email at: Richard_Brown@Brown.edu. Susan E. Ramsey is with
Rhode Island Hospital/Brown Medical School, Providence, RI. Christopher
W. Kahler and Peter M. Monti are with the Brown Center for Alcohol and
Addiction Studies/Brown Medical School, Providence, RI. Kathleen M Palm
is with Clark University, Worcester, MA. David Abrams is with the American
Legacy Foundation. Maryella Dubreuil is with Butler Hospital, Providence,
Received: May 4, 2010. Revision: August 21, 2010.
*This study was supported by National Institute on Alcohol Abuse and
†Correspondence may be sent to Richard A. Brown at the above address
1995; Kandel et al., 2001; Kessler et al., 2005; Regier et al.,
1990), with particularly high rates of comorbidity among
treatment-seeking individuals with alcohol dependence
(Grant, 1997; Lynskey, 1998). In addition to the high rates
of actual depressive disorders, the majority of patients en-
tering treatment for an alcohol use disorder report clinically
signifi cant levels of depressive symptoms that may, in part,
be the result of the physiological effects of alcohol (Dorus
et al., 1987; Nakamura et al., 1983; Schuckit et al., 1997).
Yet, studies on the association between major depression and
HE CO-OCCURRENCE OF DEPRESSION and alcohol
use disorders is well documented (Grant and Harford,
alcoholism treatment outcomes have been decidedly mixed,
with some fi nding improved outcomes among those with
comorbid depression and alcohol dependence (Charney et
al., 1998), some fi nding poorer outcomes (Greenfi eld et al.,
1998; Kranzler et al., 1996), many fi nding no association
(Booth et al., 1991; Miller et al., 1997; Sellman and Joyce,
1996), and others fi nding improved outcomes in depressed
females and poorer outcomes in depressed males (Hessel-
brock, 1991; Rounsaville et al., 1987). Still, data suggest
that compared with those who solely have either depression
or alcohol dependence, patients with co-occurring depression
and alcohol dependence tend to report more severe sympto-
mology and poorer outcomes (Greenfi eld et al., 1998; Kran-
zler et al., 1996; Rounsaville et al., 1987; Thase et al., 2001).
Given the potentially deleterious effects of depression
on alcohol treatment outcomes and the demonstrated ef-
fi cacy of cognitive-behavioral treatment for depression
(CBT-D; Brown and Lewinsohn, 1984b; Jarrett and Rush,
1994; Lewinsohn et al., 1984), two studies have examined
the effi cacy of CBT-D for reducing alcohol use in patients
with comorbid alcohol dependence and depression. One
early study reported positive fi ndings using this approach,
but more treatment contact was provided with CBT-D ver-
BROWN ET AL. 287
sus the standard comparison treatment (Monti et al., 1990;
Turner and Wehl, 1984). In a subsequent pilot study that cor-
rected this design fl aw (Brown et al., 1997), we evaluated the
comparative effi cacy of adding CBT-D versus a relaxation
training control (RTC) to standard partial hospital alcohol
treatment in a sample of 35 alcohol-dependent individuals
with elevated levels of depressive symptoms. As expected,
depressive symptoms for all participants decreased with
abstinence. However, fi ndings revealed that CBT-D patients
had signifi cantly greater reductions in somatic depressive
symptoms and depressed and anxious mood after treatment
than RTC patients. Further, preliminary analyses suggested
that decreases in somatic depressive symptoms served
to mediate the relationship between treatment condition
and drinking outcomes. Patients receiving CBT-D had a
greater percentage of days abstinent but not greater overall
abstinence or fewer drinks per day during the fi rst 3-month
follow-up. However, during the second 3 months of the
6-month follow-up period, CBT-D patients had signifi cantly
better alcohol use outcomes compared with RTC patients on
total abstinence (47% vs. 13%, respectively), percentage of
days abstinent (90.5% vs. 68.3%, respectively), and drinks
per day (0.46 vs. 5.71, respectively). Although the results of
this study were very promising, the sample size was modest,
and the follow-up extended to only 6 months. Thus, a larger-
scale clinical trial with a longer follow-up was warranted to
evaluate the effi cacy of CBT-D for alcohol dependence and
elevated depressive symptoms.
The present study represents an attempt to replicate the
fi ndings from the Brown et al. (1997) study with a larger
sample size. In the present study, the two treatment protocols
were the same as in the prior study. However, in the prior
study, treatment sessions were delivered over 8 consecutive
days while participants were engaged in partial hospital
treatment for alcohol dependence, whereas in the present
study, almost all eight individual treatment sessions were
delivered on an outpatient basis over 6 weeks. This was
because the average length of stay in the partial hospital
program from which participants were recruited decreased
considerably (from M = 21.2, SD = 4.5 to M = 3.9, SD =
2.3 calendar days) in the intervening time between the two
We hypothesized that the use of CBT-D as an adjunct to
partial hospital alcohol treatment would result in decreased
levels of depressive symptoms and in reduced quantity and
frequency of alcohol use over a 12-month follow-up relative
to the use of a relaxation control. Although we anticipated a
reduction in depressive symptoms upon a period of contin-
ued abstinence regardless of treatment condition, we hypoth-
esized that participants in the CBT-D condition would report
greater reductions relative to those in the control condition.
Additionally, we expected that the predicted reductions in
drinking in the CBT-D group would be mediated by de-
creases in depressive symptoms.
cedures as outlined for multiple regression by Cohen (1988),
with α = .05 and power = .80. Data from the pilot investiga-
tion (Brown et al., 1997) served as the source for estimates
of effect size. Participants were 166 men and women recruit-
ed from the intensive day and evening partial hospital pro-
gram, a component of Alcohol and Drug Treatment Services
(ADTS) at Butler Hospital, a private, nonprofi t hospital in
Providence, Rhode Island, providing treatment of psychiat-
ric and substance use disorders. Patients, ages 18-65 years,
were included in the study provided they met Diagnostic
and Statistical Manual of Mental Disorders, Fourth Edition
(DSM-IV; American Psychiatric Association, 1994), criteria
for alcohol dependence as determined by the patient version
of the Structured Clinical Interview for DSM-IV (SCID-P;
First et al., 1995) and had a Beck Depression Inventory
(BDI; Beck et al., 1961) total score of 15 or greater. Partici-
pants are referred to throughout this paper as patients with
alcohol dependence and elevated depressive symptoms to
avoid incorrectly suggesting that participants met diagnostic
criteria for a depressive disorder. Patient exclusion criteria
were current suicidality or homicidality, history of psychotic
disorder or current psychotic symptoms, current DSM-IV
diagnosis of opioid dependence, current DSM-IV diagnosis
of bipolar disorder, and marked organic impairment.
Following admission to the ADTS partial hospital pro-
gram, patients were informed about the study and consent
was obtained. Potential participants were then given a
screening interview to determine eligibility. Eligible patients
who consented to participate completed the pre-treatment
assessment battery and were assigned to receive either indi-
vidual CBT-D or the individual RTC. Potential participants
were recruited in sequential, nonoverlapping cohorts, with
a 1-week break in recruitment between cohorts to minimize
the possibility of treatment contamination should participants
compare treatment manuals and procedures. The number
and size of the cohorts were determined in advance, and
treatment assignment was randomly selected from the fi xed
pool of possible assignments. Consecutive admissions were
recruited into the study until there were 15 participants in the
Participants for the study were recruited from November
1997 through March 2000, and follow-ups were conducted
from December 1997 through April 2001. The fl ow of
participants through all phases of the study is presented in
Figure 1. During the recruitment phase, there were 1,499
admissions to the intensive day and evening programs. Of
those, 511 were not alcohol dependent, 203 did not have
BDI scores of 15 or greater, and 117 did not meet the age
inclusion criterion. Others meeting exclusion criteria were
Sample size for this study was determined following pro-
288 JOURNAL OF STUDIES ON ALCOHOL AND DRUGS / MARCH 2011
FIGURE 1. Flow of participants through the study. RTC = relaxation training control; CBT-D = cognitive behavioral treatment for depression.
BROWN ET AL. 289
for opiate dependence (n = 171), bipolar disorder (n = 47),
history of psychotic disorder or current psychotic symptoms
(n = 25), organic impairment (n = 5), or suicidality (n = 4).
Other reasons for noninclusion in the study were early pro-
gram discharge (n = 23), washout period (n = 32), unable
to recruit because of research staff unavailability (n = 16),
participating in another study (n = 9), and other reasons (n =
6). A total of 153 patients refused to participate in the study
before the pre-assessment battery, and 11 patients dropped
out of the study during assessment. There were no signifi cant
differences between those who refused to participate and
actual participants in age (M = 40.7 vs. 40.8 years, respec-
tively) or gender (38.6% vs. 33% female, respectively). The
most frequently cited reasons for refusal to participate in the
study included “no time” (44.4%), “not interested” (21.6%),
and “live too far away” (9.2%).
Women comprised 33% (n = 55) of the sample, and
the mean age of participants was 40.8 years (SD = 9.1).
Participants reported a mean education of 13.4 years (SD =
2.7). The sample was 93.4% White (not of Hispanic origin),
2.4% African American, 1.8% Hispanic/Latino, and 2.4%
other ethnicities. Marital status for the sample was 22.9%
never married, 34.9% married or living as if married, 32.5%
divorced and single, 7.2% separated, and 2.4% widowed.
Participants’ mean score on the Alcohol Dependence Scale
(Skinner and Allen, 1982) was 20.9 (SD = 7.4), and their
mean number of previous alcohol treatments was 3.3 (SD =
4.1). On the Timeline Followback (TLFB) interview (Sobell
and Sobell, 1996), participants reported that during the 6
months before treatment they were abstinent 38.8% (SD
= 34.1) of possible drinking days (i.e., days not in jail or
residential treatment) and consumed a mean of 13.3 (SD =
9.3) standard drinks per drinking day. Forty-eight partici-
pants (28.9%) met criteria for dependence on one or more
drugs (30 for cocaine, 18 for cannabis, 6 for sedatives, and
4 for stimulants). Because of the inclusion criteria, all par-
ticipants had a BDI score of 15 or greater; the mean BDI
was 25.6 (SD = 8.0). The mean on the Modifi ed Hamilton
Rating Scale for Depression (MHRSD; Miller et al., 1985)
was 17.2 (SD = 8.9). As reported elsewhere (Ramsey et al.,
2004), 54 of the 166 participants did not meet criteria for
current depressive disorder, and 17 met criteria for current
major depressive disorder at baseline. Of the 95 participants
who met criteria for a substance-induced mood disorder
with depressive features at baseline, 23 continued to meet
full criteria for a major depressive disorder after a 4-week
abstinence period (Ramsey et al., 2004).
which time participants had a minimum of 2-3 days of ab-
stinence from alcohol. Follow-up interviews were conducted
at 6 weeks, 3 months, 6 months, and 12 months following
An assessment battery was administered pretreatment, at
recruitment into the study, with completion rates of 86%,
93%, 95%, and 93%, respectively. Before each interview,
participants provided a breath sample to establish that they
were alcohol negative (blood alcohol concentration < .01).
Interviews were rescheduled on occasions when a breath
sample indicated that a participant was alcohol positive.
Every effort was made to ensure that the research assistants
who conducted the interviews were unaware of treatment
Diagnostic and descriptive measures. Participants pro-
vided demographic and background information such as age,
gender, years of education, and number of previous alcohol
treatments. The lifetime and current prevalence of DSM-III-
R (American Psychiatric Association, 1987) alcohol use and
affective disorders were determined by the SCID-P (Spitzer
et al., 1989). The severity of alcohol dependence was as-
sessed using the Alcohol Dependence Scale (Skinner and
Dependent measures of depressive symptoms. The BDI
(Beck et al., 1961) and the interviewer-based MHRSD (Mill-
er et al., 1985) were used to assess depressive symptoms.
Both self-report and observer-rated measures of depressive
symptoms were used because of potential discordance be-
tween the two types of measures (Enns et al., 2000).
Dependent measures of alcohol use. The TLFB interview
(Sobell et al., 1980) was used to assess drinking frequency
and quantity at baseline and during the follow-up intervals.
The TLFB was administered at baseline for the 180 days be-
fore admission and at each follow-up interval for the period
since its last administration. TLFB data include the number
of days abstinent from alcohol, amount consumed during
each drinking occasion, and the number of days in jail or
in residential treatment. One family member or close friend
(here termed signifi cant other) completed a structured inter-
view concerning the patient’s drinking behavior at 3 months,
6 months, and 12 months following study enrollment.
pital treatment program, which is an abstinence-oriented
group treatment that provides treatment daily from 9:00
A.M. to 3:30 P.M., is theoretically grounded in a cognitive
social learning model, and includes strong encouragement
for participation in 12-step programs. The program has been
described in greater detail elsewhere (McCrady et al., 1985)
and includes psychoeducation about triggers, functional
analysis, drink/drug refusal skills, and medication manage-
ment. The average length of stay at the ADTS partial hospital
program was 3.9 calendar days; thus, the CBT-D and RTC
sessions were primarily aftercare treatments. Both the CBT-
D and RTC conditions were administered in eight, 45-min-
ute individual treatment sessions that were delivered over a
6-week period on an outpatient basis after discharge from the
All participants received the standard ADTS partial hos-
290 JOURNAL OF STUDIES ON ALCOHOL AND DRUGS / MARCH 2011
ADTS partial hospital program. Detailed therapist manuals
were used to ensure standardized delivery of content in both
conditions. Doctoral-level psychologists delivered the treat-
ment. Therapists were crossed with treatment assignment,
such that each therapist provided treatment to patients in
both treatment conditions. The senior author (RAB) trained
the therapists and conducted weekly group supervision
sessions throughout the study to ensure standardization of
Cognitive-behavioral treatment for depression. The Cop-
ing With Depression course (Brown and Lewinsohn, 1984a,
1984b), modifi ed for use with alcohol-dependent patients,
served as the basis for the CBT-D treatment. CBT-D (Brown
et al., 1997; Brown and Ramsey, 2000) is an 8-session,
individual treatment for depression, incorporating training
in several depression-relevant skills, including daily mood
monitoring, pleasant activities, constructive thinking, and
social skills and assertiveness. This treatment differed from
the standard ADTS partial hospital program in that CBT-D
focused on depressive symptoms and the relationship be-
tween depression and drinking, whereas the ADTS program
emphasized changing specifi c drinking-related thoughts and
Relaxation training control. An RTC condition was used
to equate for contact time and to provide a more stringent
test of the incremental effi cacy of the CBT-D treatment.
Participants in this condition received eight individual ses-
sions that consisted of training and practice in meditative
and deep-breathing techniques (Benson, 1975), Progressive
Muscle Relaxation (Jacobson, 1929), and guided imagery.
For more details, see Brown et al. (1997). These techniques
were presented to patients as viable alternatives to drinking
and as coping skills to help decrease feelings of stress and
Treatment assignment and compliance
ment, 161 attended an initial treatment session (80 in RTC
and 81 in CBT-D) where they learned of their treatment as-
signment. Treatment conditions did not differ signifi cantly
on demographics, drinking variables, or depression-related
variables. Analyses of outcomes are based on this intent-
to-treat sample of 161. The number of participants who
completed baseline but did not attend a treatment session (n
= 5) was so small that meaningful comparisons with treat-
ment attendees could not be made. Attendance at treatment
sessions also did not differ by condition, t(159) = 1.45, p =
.15. Out of eight sessions, participants attended an average
of 7.2 (SD = 2.0) in the RTC condition and 6.7 sessions (SD
= 2.4) in the CBT-D condition. Sixty-four of 83 participants
(77.1%) in RTC completed all eight sessions, compared with
Of the 166 participants who completed a baseline assess-
57 of 83 (68.7%) in CBT-D. Throughout the study, no seri-
ous adverse events were reported. At the 6-month follow-up
period, participants in the CBT-D condition were less likely
to report additional alcohol or drug treatment, χ2(1) = 6.10,
p = .01, or emotional treatment, χ2(1) = 3.42, p = .06, rela-
tive to those in RTC. It should be noted, however, that there
were disproportionate rates of missing treatment utilization
data between the two conditions, with RTC missing 7.5%
and CBT-D missing 20.9%.
Validity of self-reports of abstinence
of abstinence at the 3-, 6-, and 12-month follow-ups. Valid
signifi cant-other reports (confi dent or very confi dent about
their knowledge of participant’s drinking) were obtained
for 68 of 76 (89%) participants reporting abstinence at 3
months, 48 of 52 (92%) reporting abstinence from 4 to 6
months, and 43 of 49 (88%) reporting abstinence from 7
to 12 months. Of these valid reports, cases in which the
participant reported abstinence and the signifi cant other
reported that the participant was drinking were rare: 2 of 68
at 3 months, 3 of 48 at 6 months, and 2 of 43 at 12 months.
These data provide support for the validity of patients’ self-
reported alcohol use over the follow-up period. In cases
of discordant reports, the worst case (i.e., drinking) was
Signifi cant-other reports were used to verify self-reports
Treatment outcome: Alcohol
day in each of the 12 months of follow-up are presented by
treatment condition in Figure 2. Before analysis, arcsine
(for percentage of days abstinent to correct negative skew-
ness) and square root (for drinks per drinking day to correct
positive skewness) transformations were used to improve
the distributional properties of these variables. Hierarchi-
cal linear modeling (HLM; Bryk and Raudenbush, 1987)
was conducted in SAS (SAS Institute Inc., Cary, NC) us-
ing PROC MIXED to test between-group differences in
individual growth curves over the 12-month follow-up. We
conducted analyses using all available data. This allowed us
to include 97.5% (n = 157) of the sample in the drinking
outcome analyses. We initially modeled the random effects
of therapist and randomization cohort as separate levels in
the model; however, both of these variance terms were non-
signifi cant and were therefore dropped.
In each analysis, we included as covariates in the Level
2 model the corresponding variable at baseline as well as
gender, baseline BDI score, and history of independent
major depressive disorder. Model fi t indices indicated that
individual growth trajectories (Level 1 model) were best
modeled using both linear and quadratic slope terms. The
HLMs predicting percentage of days abstinent and drinks
The percentage of days abstinent and drinks per drinking
BROWN ET AL. 291
FIGURE 2. Percentage of days abstinent and drinks per drinking day by treatment condition and time. RTC = relaxation training control; CBT-D = cognitive
behavioral treatment for depression.
292 JOURNAL OF STUDIES ON ALCOHOL AND DRUGS / MARCH 2011
FIGURE 3. Depressive symptoms by treatment condition and time. RTC = relaxation training control; CBT-D = cognitive behavioral treatment for depression.
BROWN ET AL. 293
per drinking day across the 12 months of follow-up indi-
cated the treatment conditions did not differ signifi cantly
on the mean levels of either variable (ps = .92 and .83,
respectively). There also were no main effects for gender,
baseline BDI score, or history of independent major de-
pressive disorder, nor did treatment condition interact with
any of these variables. Although there were signifi cant,
negative linear effects and signifi cant positive quadratic
effects of time, treatment conditions did not differ signifi -
cantly in these parameters. Analyses were repeated exam-
ining only participants who completed all eight treatment
sessions in either condition. Again, differences between
conditions were nonsignifi cant.
Treatment outcome: Depressive symptoms
the 6-week, 3-month, 6-month, and 12-month follow-ups are
presented in Figure 3. Before analyses, both variables were
subjected to square-root transformation to correct positive
skewness. HLM was used to test for differences in individual
growth curve parameters in depressive symptoms between
treatment conditions. All available data were used in these
analyses, providing a sample size of 157. The cohort level
of the model was dropped because that variance component
was nonsignifi cant.
In the HLMs predicting BDI and MHRSD scores, we
covaried for the corresponding variable at baseline, as well
as gender, history of independent major depressive disorder,
and time. Based on model fi t indices, we modeled time in
the Level 1 model using the log-transformation of time of
assessment, rather than using both a linear and quadratic
effect of time. The Level 2 (between-subjects) parameter
estimate for the intercept indicated the treatment conditions
did not differ signifi cantly on mean levels of either the BDI
or MHRSD across the follow-ups (ps = .31 and .27, respec-
tively). There also were no signifi cant main effects for gen-
der. Having a history of past or current independent major
depressive disorder was associated with greater BDI scores
across the follow-ups (B = 0.47, SE = 0.22, p =.04), although
the effect did not reach signifi cance with the MHRSD (B =
0.33, SE = 0.20, p = .10). Treatment condition did not inter-
act signifi cantly with any of the other covariates in predicting
mean levels of depressive symptoms.
The Level 2 (between-subjects) parameter estimate for
the log of time indicated that MHRSD scores decreased sig-
nifi cantly over time (B = -0.13, SE = 0.05, p =.01), whereas
BDI scores did not (B = -0.08, SE = 0.06, p = .16). The ef-
fect of time (i.e., the rate of change in depressive symptoms)
differed by condition for BDI scores, with the effect of time
being signifi cantly more negative in RTC (B = -0.35, SE =
0.11, p = .001). Given the different rates of change in BDI
scores over time, we conducted model-based estimates of
treatment effects at each follow-up. These analyses indicated
Mean BDI and MHRSD scores by treatment condition at
that, compared with participants receiving RTC, those in
CBT-D scored lower on the BDI at the 6-week follow-up (B
= -0.56, SE = 0.23, p = 02). However, from that point for-
ward, RTC showed greater reductions in BDI scores relative
to CBT-D such that the conditions did not differ signifi cantly
on BDI scores at any later follow-ups (ps > .20). The effect
of treatment condition on rate of change was nonsignifi cant
for MHRSD scores (p = .11), although results followed a
similar pattern. Analyses were repeated examining only
participants who completed all eight treatment sessions in
either condition. Again, differences between conditions were
addition to standard alcohol treatment was associated with
positive outcomes among alcohol-dependent individuals
with elevated depressive symptoms and how those outcomes
compared with RTC plus standard treatment. Although we
anticipated that both conditions would demonstrate improve-
ment in depressive symptoms, in part related to decreased
physiological effects of alcohol during a period of absti-
nence, we hypothesized that participants receiving CBT-D
would report a greater change in depressive and alcohol use
outcomes relative to the control condition. Indeed, both treat-
ment conditions demonstrated improvements in depressive
and drinking outcomes over time. Compared with RTC, the
CBT-D condition had signifi cantly lower levels of depressive
symptoms, as measured by the BDI, at the 6-week follow-
up. However, this effect was inconsistent because there were
no differences in the MHRSD between conditions at the
6-week follow-up and there were no signifi cant differences
in depressive symptoms at any other follow-up. Signifi cant
between-group differences on alcohol use outcomes were
not found, either at posttreatment or during the 12-month
The current fi ndings did not replicate the positive out-
comes observed in the CBT-D condition in our previous pilot
study (Brown et al., 1997). We considered several possible
explanations for why these fi ndings were not replicated. We
do not believe that the present fi ndings were the result of
a lack of power, because this study was a well-conducted
trial with power of .80 to detect effect sizes in the small to
medium range (d = .35 to d = .40 depending on the outcome
examined) using HLM (Hedeker et al., 1999). Additionally,
this study had good treatment attendance and follow-up
rates. One possible explanation for the present fi ndings
involves consideration of the unanticipated changes to the
treatment context that occurred during the intervening time
between the two studies. During the time between the pilot
project and the current study, the average length of stay in
the partial hospital program where the studies were con-
ducted decreased from 21.2 calendar days to 3.9 calendar
This study examined whether participation in CBT-D in
294 JOURNAL OF STUDIES ON ALCOHOL AND DRUGS / MARCH 2011
days. Because of this dramatic decrease in length of stay,
the majority of the treatment sessions in the present study
were conducted on an outpatient basis (eight sessions over
6 weeks), whereas the pilot study treatment sessions were
conducted during the course of partial hospital treatment on
8 consecutive treatment days.
There are several ways that this change in treatment con-
text may have accounted for the different fi ndings between
the studies. For example, it is possible that the positive
outcomes in the pilot study were the result of a synergistic
effect in which CBT-D was more effective when presented
concurrently with CBT for alcohol dependence. Another
possibility is that because treatment in this study was pro-
vided on an outpatient basis, all participants were assured
of ongoing follow-up treatment (i.e., either CBT-D or RTC)
after discharge from partial hospital treatment, and CBT-D
outcomes (relative to RTC) could not outperform the benefi ts
related to this outpatient experience. In any event, the treat-
ment context in which CBT-D was delivered in these two
studies was quite different, and there is no way to determine
exactly how and whether these differences played a role in
the discrepant fi ndings between the studies.
One other interesting point to address is the relationships
between treatment, changes in depressive symptoms, and
alcohol outcomes. In the previous study, changes in the level
of depression mediated alcohol use outcomes. This fi nding
was not observed in the current study. These discrepant fi nd-
ings across studies could be related to the change in treat-
ment context between studies, but they also might have been
infl uenced by alternative unexamined mechanisms. There is
accumulating evidence illustrating rapid early response to
CBT for depression (Hardy et al., 2005; Tang and DeRu-
beis, 1999; Tang et al., 2005), problem drinking (Breslin et
al., 1997), and other clinical conditions (Stiles et al., 2003;
Wilson, 1999). Although there has been a debate regarding
the mechanisms responsible for early treatment response
(e.g., Hollon, 1999; Ilardi and Craighead, 1999; Tang and
DeRubeis, 1999; Wilson, 1999), the observation in the cur-
rent study of rapid early response in drinking and depressive
outcomes in both conditions suggests that treatment effects
may have been the result of nonspecifi c processes.
The current study had several limitations that merit dis-
cussion. First, the current study was designed as a replication
of a previous pilot study in which CBT-D outperformed RTC
in reducing depressive and alcohol outcomes. However, the
changes to the treatment context between the two studies
made it diffi cult to compare results, and thus the current
study was not a true replication. Second, alternative process
variables that may have been related to outcomes were not
assessed. Future studies should examine the mechanisms
through which relaxation treatment might affect alcohol
and depression outcomes in this population. Third, there are
characteristics of the current sample that may limit the gen-
eralizability of these fi ndings. A large number of potential
participants refused to participate in this study. Although
they did not differ from actual participants in gender or age,
it is possible that refusers differed from the participants in
some meaningful way such that they would have reported
different responses to the treatments in this study. Also,
given that this study was conducted in a private hospital set-
ting with well-educated, primarily White patients, caution
should be used in generalizing the current fi ndings to popu-
lations with different characteristics. A fi nal limitation of the
current study is that reliability ratings for interview data and
treatment adherence were not conducted. Although assess-
ment staff and therapists participated in on-going training
and supervision, we cannot conclude with certainty that the
interviews and treatments were delivered adequately. We do
not believe that the lack of differences in outcomes between
conditions was the result of nonadherence to the treatment
protocol. However, we are unable to rule out this possibility
based on available data.
Future studies evaluating the effi cacy of CBT-D in
alcohol-dependent individuals with elevated depressive
symptoms continue to be warranted. Future investigations
might explore the role of treatment context on the effi cacy of
CBT-D for this population. It is important to clarify whether
this treatment has incremental treatment effects only when
it works synergistically with CBT for alcohol dependence.
Also, future studies might evaluate the relative effi cacy of
CBT-D provided in a group versus individual format, given
the equivalence of these modalities demonstrated in previous
work (Brown and Lewinsohn, 1984b). Finally, future stud-
ies should examine the effi cacy of CBT-D with depressed
alcohol-dependent patients whose depressive symptoms do
not remit with abstinence. Our results support the idea that,
for many people struggling with alcohol dependence, depres-
sive symptoms remit with abstinence and may be the result
of the pharmacological effects of alcohol and/or the situa-
tional factors related to the crisis that precipitated treatment.
Interestingly, in a meta-analysis of medication response for
comorbid depression and substance dependence, Nunes and
Levin (2004) found that diagnosis of depression after 1 week
of abstinence was associated with a greater antidepressant
response, and those studies that found large effect sizes on
depressive symptoms also showed signifi cant decreases in
the quantity of substance use. It is possible that the same
kind of pattern exists for the effects of CBT-D on comorbid
depression and substance dependence. Thus, we believe that
a promising area for future research might involve the use of
CBT-D in patients whose depressive symptoms do not remit
Melanie Poisson, Andrea McGuinn, and Wendy Reinke. We also thank Peter
M. Lewinsohn for his valuable input and consultation.
We gratefully acknowledge the research assistance of Caron Francione,
BROWN ET AL. 295
American Psychiatric Association. (1987). Diagnostic and statistical manual
of mental disorders (3rd ed., rev.). Washington, DC: Author.
American Psychiatric Association. (1994). Diagnostic and statistical manual
of mental disorders (4th ed.). Washington, DC: Author.
Beck, A. T., Ward, C. H., Mendelson, M., Mock, J., & Erbaugh, J. (1961).
An inventory for measuring depression. Archives of General Psychiatry,
Benson, H. (1975). The relaxation response. New York: William Morrow.
Booth, B. M., Yates, W. R., Petty, F., & Brown, K. (1991). Patient factors
predicting early alcohol-related readmissions for alcoholics: Role of
alcoholism severity and psychiatric co-morbidity. Journal of Studies on
Alcohol, 52, 37-43.
Breslin, F. C., Sobell, M. B., Sobell, L. C., Buchan, G., & Cunningham, J.
A. (1997). Toward a stepped care approach to treating problem drink-
ers: the predictive utility of within-treatment variables and therapist
prognostic ratings. Addiction, 92, 1479-1489.
Brown, R. A., Evans, D. M., Miller, I. W., Burgess, E. S., & Mueller, T. I.
(1997). Cognitive-behavioral treatment for depression in alcoholism.
Journal of Consulting and Clinical Psychology, 65, 715-726.
Brown, R. A., & Lewinsohn, P. M. (1984a). Coping with depression: Course
workbook. Eugene, OR: Castalia Press.
Brown, R. A., & Lewinsohn, P. M. (1984b). A psychoeducational approach
to the treatment of depression: comparison of group, individual, and
minimal contact procedures. Journal of Consulting and Clinical Psy-
chology, 52, 774-783.
Brown, R. A., & Ramsey, S. E. (2000). Addressing comorbid depressive
symptomatology in alcohol treatment. Professional Psychology, Re-
search and Practice, 31, 418-422.
Bryk, A. S., & Raudenbush, S. W. (1987). Application of hierarchical linear
models to assessing change. Psychological Bulletin, 101, 147-158.
Charney, D. A., Paraherakis, A. M., Negrete, J. C., & Gill, K. J. (1998). The
impact of depression on the outcome of addictions treatment. Journal
of Substance Abuse Treatment, 15, 123-130.
Cohen, J. (1988). Statistical power for the behavioral sciences (Second edi-
tion). Hillsdale, NJ: Lawrence-Erlbaum Associates, Inc.
Dorus, W., Kennedy, J., Gibbons, R. D., & Ravi, S. D. (1987). Symptoms
and diagnosis of depression in alcoholics. Alcoholism: Clinical and
Experimental Research, 11, 150-154.
Enns, M. W., Larsen, D. K., & Cox, B. J. (2000). Discrepancies between self
and observer ratings of depression: The relationship to demographic,
clinical and personality variables. Journal of Affective Disorders, 60,
First, M. B., Spitzer, R. L., Gibbon, M., & Williams, J. B. W. (1995). Struc-
tured Clinical Interview for DSM-IV Axis I Disorders. New York: New
York State Psychiatric Institute.
Grant, B. F. (1997). The infl uence of comorbid major depression and sub-
stance use disorders on alcohol and drug treatment: Results of a national
survey. In L. S. Onken, J. D. Blaine, S. Genser, & A. M. Horton, Jr.,
(Eds.), Treatment of drug-dependent individuals with comorbid mental
disorders (NIDA Research Monograph No. 172, NIH Publication No.
97-4172, pp. 4-15). Washington, DC: Government Printing Offi ce.
Grant, B. F., & Harford, T. C. (1995). Comorbidity between DSM-IV alco-
hol use disorders and major depression: Results of a national survey.
Drug and Alcohol Dependence, 39, 197-206.
Greenfi eld, S. F., Weiss, R. D., Muenz, L. R., Vagge, L. M., Kelly, J. F.,
Bello, L. R., & Michael, J. (1998). The effect of depression on return
to drinking: A prospective study. Archives of General Psychiatry, 55,
Hardy, G. E., Cahill, J., Stiles, W. B., Ispan, C., Macaskill, N., & Barkham,
M. (2005). Sudden gains in cognitive therapy for depression: A repli-
cation and extension. Journal of Consulting and Clinical Psychology,
Hedeker, D., Gibbons, R. D., & Waternaux, C. (1999). Sample size esti-
mation for longitudinal designs and attrition: Comparing time-related
contrasts between two groups. Journal of Educational and Behavioral
Statistics, 24, 70-93.
Hesselbrock, M. N. (1991). Gender comparison of antisocial personality
disorder and depression in alcoholism. Journal of Substance Abuse, 3,
Hollon, S. D. (1999). Rapid early response in cognitive behavior therapy:
A commentary. Clinical Psychology: Science and Practice, 6, 305-309.
Ilardi, S. S., & Craighead, W. E. (1999). Rapid early response, cognitive
modifi cation, and nonspecifi c factors in cognitive behavior therapy for
depression: A reply to Tang and DeRubeis. Clinical Psychology: Science
and Practice, 6, 295-299.
Jacobson, E. (1929). Progressive relaxation. Chicago, IL: University of
Jarrett, R. B., & Rush, A. J. (1994). Short-term psychotherapy of depressive
disorders: Current status and future directions. Psychiatry, 57, 115-132.
Kandel, D. B., Huang, F. Y., & Davies, M. (2001). Comorbidity between
patterns of substance use dependence and psychiatric syndromes. Drug
and Alcohol Dependence, 64, 233-241.
Kessler, R. C., Chiu, W. T., Demler, O., Merikangas, K. R., & Walters, E.
E. (2005). Prevalence, severity, and comorbidity of 12-month DSM-IV
disorders in the National Comorbidity Survey Replication. Archives of
General Psychiatry, 62, 617-627.
Kranzler, H. R., Del Boca, F. K., & Rounsaville, B. J. (1996). Comorbid
psychiatric diagnosis predicts three-year outcomes in alcoholics: A
posttreatment natural history study. Journal of Studies on Alcohol, 57,
Lewinsohn, P. M., Antonuccio, D. O., Steinmetz, J. L., & Teri, L. (1984).
The Coping with Depression Course: A psychoeducational intervention
for unipolar depression. Eugene, OR: Castalia Press.
Lynskey, M. T. (1998). The comorbidity of alcohol dependence and affec-
tive disorders: treatment implications. Drug and Alcohol Dependence,
McCrady, B. S., Dean, L., Dubreuil, E., & Swanson, S. (1985). The problem
drinkers’ project: A programmatic application of social-learning-based
treatment. In G. A. Marlatt & J. R. Gordon (Eds.), Relapse prevention
(pp. 417-471). New York: Guilford Press.
Miller, I. W., Bishop, S., Norman, W. H., & Maddever, H. (1985). The
Modifi ed Hamilton Rating Scale for Depression: reliability and validity.
Psychiatry Research, 14, 131-142.
Miller, N. S., Hoffmann, N. G., Ninonuevo, F., & Astrachan, B. M. (1997).
Lifetime diagnosis of major depression as a multivariate predictor of
treatment outcome for inpatients with substance use disorders from
abstinence-based programs. Annals of Clinical Psychiatry, 9, 127-137.
Monti, P. M., Abrams, D. B., Binkoff, J. A., Zwick, W. R., Liepman, M.
R., Nirenberg, T. D., & Rohsenow, D. J. (1990). Communication skills
training, communication skills training with family and cognitive be-
havioral mood management training for alcoholics. Journal of Studies
on Alcohol, 51, 263-270.
Nakamura, M. M., Overall, J. E., Hollister, L. E., & Radcliffe, E. (1983).
Factors affecting outcome of depressive symptoms in alcoholics. Alco-
holism: Clinical and Experimental Research, 7, 188-193.
Nunes, E. V., & Levin, F. R. (2004). Treatment of depression in patients
with alcohol or other drug dependence: a meta-analysis. Journal of the
American Medical Association, 291, 1887-1896.
Ramsey, S. E., Kahler, C. W., Read, J. P., Stuart, G. L., & Brown, R. A.
(2004). Discriminating between substance-induced and independent
depressive episodes in alcohol dependent patients. Journal of Studies
on Alcohol, 65, 672-676.
Regier, D. A., Farmer, M. E., Rae, D. S., Locke, B. Z., Keith, S. J., Judd,
L. L., & Goodwin, F. K. (1990). Comorbidity of mental disorders with
alcohol and other drug abuse. Results from the Epidemiologic Catch-
296 JOURNAL OF STUDIES ON ALCOHOL AND DRUGS / MARCH 2011
ment Area (ECA) Study. Journal of the American Medical Association,
Rounsaville, B. J., Dolinsky, Z. S., Babor, T. F., & Meyer, R. E. (1987).
Psychopathology as a predictor of treatment outcome in alcoholics.
Archives of General Psychiatry, 44, 505-513.
Schuckit, M. A., Tipp, J. E., Bergman, M., Reich, W., Hesselbrock, V. M.,
& Smith, T. L. (1997). Comparison of induced and independent major
depressive disorders in 2,945 alcoholics. American Journal of Psychia-
try, 154, 948-957.
Sellman, J. D., & Joyce, P. R. (1996). Does depression predict relapse in
the 6 months following treatment for men with alcohol dependence?
The Australian and New Zealand Journal of Psychiatry, 30, 573-578.
Skinner, H. A., & Allen, B. A. (1982). Alcohol dependence syndrome: Mea-
surement and validation. Journal of Abnormal Psychology, 91, 199-209.
Sobell, L. C., & Sobell, M. B. (1996). Timeline followback: A calendar
method for assessing alcohol and drug use. Toronto, Canada: Addiction
Sobell, M. B., Maisto, S. A., Sobell, L. C., Cooper, A. M., Cooper, T. C.,
& Sanders, B. (1980). Developing a prototype for evaluating alcohol
treatment effectiveness. In L. C. Sobell, M. B. Sobell, & E. Ward (Eds.),
Evaluating alcohol and drug abuse treatment effectiveness: Recent
advances. New York: Pergamon Press.
Spitzer, R. L., Williams, J. B. W., Gibbon, M., & First, M. B. (1989). User’s
guide for the Structured Clinical Interview for DSM-III-R. New York:
New York Psychiatric Institute.
Stiles, W. B., Leach, C., Barkham, M., Lucock, M., Iveson, S., Shapiro, D.
A., . . . Hardy, G. E. (2003). Early sudden gains in psychotherapy under
routine clinic conditions: Practice-based evidence. Journal of Consulting
and Clinical Psychology, 71, 14-21.
Tang, T. Z., & DeRubeis, R. J. (1999). Sudden gains and critical sessions in
cognitive-behavioral therapy for depression. Journal of Consulting and
Clinical Psychology, 67, 894-904.
Tang, T. Z., DeRubeis, R. J., Beberman, R., & Pham, T. (2005). Cognitive
changes, critical sessions, and sudden gains in cognitive-behavioral
therapy for depression. Journal of Consulting and Clinical Psychology,
Thase, M. E., Salloum, I. M., & Cornelius, J. D. (2001). Comorbid alcohol-
ism and depression: Treatment issues. The Journal of Clinical Psychia-
try, 62 (Suppl. 20), 32-41.
Turner, R. W., & Wehl, C. K. (1984). Treatment of unipolar depression in
problem drinkers. Advances in Behaviour Research and Therapy, 6,
Wilson, G. T. (1999). Rapid response to cognitive behavior therapy. Clinical
Psychology: Science and Practice, 6, 289-292.