RESEARCH REPORT doi:10.1111/j.1360-0443.2008.02444.

Computer-based psychological treatment for

comorbid depression and problematic alcohol and/or
cannabis use: a randomized controlled trial of clinical
efficacy

Frances J. Kay-Lambkin1, Amanda L. Baker2, Terry J. Lewin3 & Vaughan J. Carr4

Centre for Brain and Mental Health Research, University of Newcastle, Newcastle, NSW, Australia, 2300 and National Drug and Alcohol Research Centre,
University of New South Wales, Sydney NSW 2052, Australia,1 Centre for Brain and Mental Health Research, University of Newcastle, Newcastle, NSW 2300,
Australia,2 Centre for Brain and Mental Health Research and Hunter New England Mental Health, Newcastle, NSW 2300, Australia3 and Centre for Brain and
Mental Health Research, University of Newcastle, Newcastle, NSW 2300, Australia and Schizophrenia Research Institute, Sydney, Australia4

ABSTRACT

Aims To evaluate computer- versus therapist-delivered psychological treatment for people with comorbid depression
and alcohol/cannabis use problems. Design Randomized controlled trial. Setting Community-based participants in
the Hunter Region of New South Wales, Australia. Participants Ninety-seven people with comorbid major depression
and alcohol/cannabis misuse. Intervention All participants received a brief intervention (BI) for depressive symptoms
and substance misuse, followed by random assignment to: no further treatment (BI alone); or nine sessions of moti-
vational interviewing and cognitive behaviour therapy (intensive MI/CBT). Participants allocated to the intensive
MI/CBT condition were selected at random to receive their treatment ‘live’ (i.e. delivered by a psychologist) or via a
computer-based program (with brief weekly input from a psychologist). Measurements Depression, alcohol/cannabis
use and hazardous substance use index scores measured at baseline, and 3, 6 and 12 months post-baseline assessment.
Findings (i) Depression responded better to intensive MI/CBT compared to BI alone, with ‘live’ treatment demonstrat-
ing a strong short-term beneficial effect which was matched by computer-based treatment at 12-month follow-up; (ii)
problematic alcohol use responded well to BI alone and even better to the intensive MI/CBT intervention; (iii) intensive
MI/CBT was significantly better than BI alone in reducing cannabis use and hazardous substance use, with computer-
based therapy showing the largest treatment effect. Conclusions Computer-based treatment, targeting both depres-
sion and substance use simultaneously, results in at least equivalent 12-month outcomes relative to a ‘live’
intervention. For clinicians treating people with comorbid depression and alcohol problems, BIs addressing both issues
appear to be an appropriate and efficacious treatment option. Primary care of those with comorbid depression and
cannabis use problems could involve computer-based integrated interventions for depression and cannabis use, with
brief regular contact with the clinician to check on progress.

Keywords Comorbidity, computer-based treatment, depression, substance use.

Correspondence to: Frances Kay-Lambkin, Centre for Brain and Mental Health Research, University of Newcastle, PO Box 833, Newcastle, NSW, 2300,
Australia. E-mail: frances.kaylambkin@newcastle.edu.au
Submitted 6 June 2008; initial review completed 29 September 2008; final version accepted 20 October 2008

INTRODUCTION epidemiology and characteristics of people with comor-

Depression and alcohol/other drug (AOD) use disorders
are two of the three most common and disabling mental
disorders [1]. These disorders also co-occur with consid-
erable frequency, a phenomenon referred to as ‘comor-
bidity’ [2]. Although much is known about the
bid disorders, much less is known about effective treat-
ment strategies to assist this increasingly prevalent
clinical group, who often do not access treatment for their
conditions [3].
Computers and the internet offer a potential solution
[4]. Treatment via these modalities may enable treatment

© 2009 The Authors. Journal compilation © 2009 Society for the Study of Addiction Addiction, 104, 378–388
 Computerized psychological treatment for comorbidity
access by groups typically disadvantaged due to geo-
graphical isolation, socio-economic status, stigma and
skill shortage of therapists confident in addressing
comorbidity. Cognitive–behaviour therapy (CBT) is the
psychological treatment of choice for both depression
and AOD use problems, and has been adapted for delivery
via computer programs and/or the internet for the follow-
ing conditions: depression [5–7]; anxiety [8,9]; smoking
cessation [10]; problem drinking [11,12]; and eating dis-
orders [13]. Preliminary evidence from these studies sug-
gests potential benefits [14,15]. For example, a recent
meta-analysis of the efficacy of internet-based CBT for
symptoms of depression and anxiety indicated that,
across 12 randomized controlled trials involving 2334
participants, internet-based CBT was associated with
moderate to large mean effect sizes (i.e. 0.40–0.60) rela-
tive to treatment as usual, waiting-list or information
controls [16]. More recently, when combined with
minimal therapist contact (e.g. e-mails supporting the
computerized content), computer-based CBT has been
associated with similar improvements in key outcomes
as a ‘live’ therapist-delivered intervention [17]. The
National Institute for Clinical Excellence (NICE) in the
United Kingdom suggests that computerized CBT may be
of value in clinical settings, especially in managing con-
ditions such as anxiety and depression [18]. Although it
is acknowledged that treatment would need to encom-
pass the comorbidities with which people experiencing
depression (and anxiety) increasingly encounter,
computer-based interventions have not yet been devel-
oped or tested for comorbid problems, such as concurrent
major depression and problematic AOD use.
We attempted to address the gap in evidence by
conducting a randomized controlled trial of computer-
versus therapist-delivered intensive motivational inter-
view (MI)/CBT (SHADE therapy: Self-Help for Alcohol
and other drug use and Depression) versus BI alone for
people with comorbid depression and problematic
alcohol or cannabis use. It was hypothesized that: (i) par-
ticipants in the intensive MI/CBT conditions (computer-
and therapist-delivered SHADE) would produce superior
improvements in depression and alcohol or cannabis use
relative to BI alone; and (ii) the differences in alcohol or
cannabis and depression-related outcomes for computer-
and therapist-delivered SHADE would be compared, and
not differ by more than 0.25 standard deviation (SD;
effect size) units.
METHODS
Study participants and location
The study was conducted in the Hunter Region of NSW,
Australia [19]. Referrals to the project were sought from
© 2009 The Authors. Journal compilation © 2009 Society for the Study of Addiction
379
AOD, mental health and primary health-care settings.
Participants were also drawn from the general commu-
nity, in response to advertising through the local televi-
sion and print media. The eligibility criteria were: (i) a
score of 17 or greater on the Beck Depression Inventory II
(BDI-II, 20); (ii) life-time diagnosis of major depressive
disorder, as confirmed by the Structured Clinical Inter-
view for DSM-IV (SCID-RV, [21]); (iii) current problematic
AOD use—alcohol consumption above recommended
drinking levels in Australia (four standard drinks per day
for men or two standard drinks per day for women, [22])
or at least weekly use of cannabis; (iv) absence of a brain
injury, organic brain disease and/or significant cognitive
impairment; (v) aged over 16 years; and (vi) ability to
understand English. Participants could be using other
substances, but must have met threshold for problematic
use of alcohol or cannabis at entry to the study.
In total, the project received 169 referrals over an
18-month period, with 97 people (57%) meeting study
criteria and commencing the project (see Fig. 1). All
participants were volunteers and received up to AU$20
reimbursement of travel expenses at each assessment.
Treatment was provided free of charge.
Study design
Following provision of informed consent, eligible partici-
pants completed face-to-face baseline assessment with a
research clinician (120 minutes) and commenced the
treatment phase of the study, which started with a brief
intervention (BI). At the conclusion of the BI, partici-
pants were assigned randomly to receive: no further
treatment, or nine further sessions of SHADE therapy
over 3 months delivered by a ‘live’ psychologist or by a
computer program (with brief 10–15-minute weekly
psychologist input). A randomization list was generated
independently and linked to a unique participant identi-
fication code (i.e. 1–120). Treatment allocations were
transferred from this list by an administrative assistant
and concealed in individual envelopes labelled with the
relevant participant code. Neither of these processes was
conducted by personnel involved with the assessment or
treatment phases of the study. Prior to the BI session, the
research clinicians were issued with a new randomiza-
tion envelope by the administrative assistant, which dis-
played the participant number on the outside of the
envelope with the treatment allocation sealed inside. The
envelope was opened by the participant at the conclusion
of the BI session. Research clinicians were blind to treat-
ment allocation until the conclusion of the BI. A per-
muted block randomization approach was used so that
the distribution of participants across treatment condi-
tions could be maintained regardless of the final sample
size.
Addiction, 104, 378–388
380 Frances J. Kay-Lambkin et al.

Assessed for eligibility (n=169)

Excluded (n=72)
Not meeting inclusion criteria (n=44)
Refused to participate (n=19)
Uncontactable (n=9)

Eligible to enter trial (n=97)

Baseline assessment with therapist (n=97)

Brief Intervention (1 session with therapist, n=97)

Random allocation (n=97)

9-sessions of therapist- 9-sessions of computer- No further treatment

delivered CBT delivered CBT
Allocated to intervention Allocated to intervention Allocated to intervention
(n=35) (n=32) (n=30)
Received all sessions Received all sessions
(n=19) (n=15)

Completed post-treatment follow-up (15-weeks post-initial, n=82)

Refused to participate (n=3)
Uncontactable (n=12)

Completed 6-month follow-up (26-weeks post-initial, n=79)

Refused to participate (n=4)
Uncontactable (n=14)

Completed 12-month follow-up (52-weeks post-initial, n=82)

Refused to participate (n=4)
Uncontactable (n=11)

Analyzed
Baseline data (n=97)
Outcome data (n=67) – those who completed all assessments
Figure 1 Flow of participants through
the study

At the conclusion of the treatment period all partici-
pants, regardless of treatment completion, met with an
independent research clinician, blind to treatment alloca-
tion, to complete follow-up assessments. Follow-up
occurred at 3, 6 and 12 months after the baseline.
as a general non-confrontational approach to discussions
regarding making and maintaining changes [27]. MI
and CBT are complementary clinical approaches with
evidence of benefit for both depression and AOD use
problems [28].

Interventions
A harm minimization approach to reducing alcohol and
cannabis use was emphasized, with participants choos-
ing their therapy goals [23–25]. CBT strategies were inte-
grated by allowing for recognition and exploration of the
relationship between depressive symptoms and alcohol/
cannabis use problems, including how each condition
may be exacerbated by the other [26]. MI was used
throughout treatment, with early sessions integrating
specific techniques with CBT strategies (e.g. decisional
balance, developing change plans), and in later sessions
BI (control)
This one-session manualized intervention was delivered
to all participants face to face and comprised rapport
building, case formulation, feedback from assessment, MI
[29], brief advice to reduce alcohol/cannabis use and self-
help material for depression and alcohol/cannabis use
problems. Participants allocated subsequently to receive
further therapy were asked to complete a daily mood/
alcohol/cannabis monitoring task over the coming week
[30,31].

© 2009 The Authors. Journal compilation © 2009 Society for the Study of Addiction Addiction, 104, 378–388
 Computerized psychological treatment for comorbidity
Therapist-delivered SHADE intensive therapy
Treatment consisted of 10 individual sessions of therapy,
1 week apart, delivered by a psychologist, including the BI
as session 1. SHADE therapy incorporated MI and CBT
components [30–35], and each session followed a
detailed treatment manual.
Computer-delivered SHADE intensive therapy
The content of the therapist- and computer-delivered
SHADE therapy was identical, with the latter including
interactive components such as video demonstrations,
voice-overs and in-session exercises. Participants were
instructed to complete the nine sessions in sequence, 1
week apart, as per the therapist-delivered intervention,
and attended the research centre to access the computer
program. Following completion of each SHADE comput-
erized module, the research clinician met with the
participant for a manualized 10–15-minute ‘check-in’
session, which included: a review of homework activities;
the development of a plan for completing homework; a
brief suicide risk and mood assessment; and confirmation
of the next appointment.
Treatment fidelity
All treatment sessions and brief check-in sessions were
tape-recorded, with a 20% sample being selected ran-
domly for evaluation by an independent psychologist.
Measures
Baseline descriptive variables
Age, gender, marital status, education levels, employ-
ment status, severity of depression and AOD use history
were measured at baseline.
Depression
The BDI-II [20] measured depressive symptoms, while
life-time and current diagnosis of major depressive disor-
der was measured using the SCID-RV [21].
AOD use
The Opiate Treatment Index (OTI, [36]) yields a score
reflecting the mean number of substance use occasions
per day for each drug class assessed for the month prior to
assessment. The alcohol and cannabis use scores on this
scale are reported. A hazardous use index score was cal-
culated, which estimated the number of day equivalents
in the previous month that participants used a range of
10 drug types at harmful levels (range 0–280). Substance
abuse and dependence was measured using the SCID-
RV[21].
© 2009 The Authors. Journal compilation © 2009 Society for the Study of Addiction
381
Statistical analyses
Data were analysed using the Statistical Package for
Social Sciences (version 14.0; SPSS, Chicago, IL, USA).
Analyses of variance were conducted on participants
who completed all four assessments, as well as on an
intention-to-treat (ITT) basis, with missing value substi-
tutions based on the last observation carried forward. As
a partial control for the number of statistical tests per-
formed, the threshold for statistical significance was set at
P < 0.01.
Baseline descriptive variables
Exploratory data analysis was conducted on the full
sample of participants (n = 97) examining basic demo-
graphic variables.
Changes in symptoms over time
The primary outcome variables were levels of depression,
alcohol and cannabis use, and the secondary outcome
variable was hazardous AOD use index scores. For depres-
sion and hazardous AOD use (n = 67), planned contrasts
from repeated-measures analyses of variance (ANOVA)
were used to examine patterns of change over time. Poly-
nomial trend contrasts (i.e. linear, quadratic and cubic
trends) were used to examine differences across assess-
ment occasions, with four contrasts used to examine dif-
ferences between groups: BI versus therapist; BI versus
computer; therapist versus computer; and BI versus com-
bined SHADE intensive intervention.
Primary outcome measures at the 12-month time-point
Logistic regression analyses were used to examine differ-
ences between groups in selected categorical outcomes at
12 months, as this time-point was considered the critical
point at which to compare the magnitude and duration
of treatment effects. A similar analysis was conducted
for alcohol (n = 41) and cannabis use (n = 43) among
follow-up completers who met baseline threshold criteria
for problematic use of each drug. Participants were
classed as ‘improved’ if they reported scores of <17 on the
BDI-II (the entry level for the study) or at least a 50%
reduction in alcohol or cannabis use. Effect size differ-
ences larger than a quarter of a standard deviation were
regarded as evidence of non-equivalence (or potential dif-
ferential clinical utility) of the treatment conditions.
RESULTS
Baseline demographic variables
Of the 97 participants, 54% (n = 52) were female, 93%
(n = 90) were Australian-born, 44% (n = 43) had never
Addiction, 104, 378–388
382 Frances J. Kay-Lambkin et al.

married and 41% (n = 40) lived with a partner. The mean
age of the sample was 35.37 years (range 18–61) and, on
average, participants had left school at 15.62 years of age
(range 12–19).
The mean BDI-II score at baseline was 31.93
(SD = 9.55, range 17–53), with 68% in the severely
depressed range (>27). Among those using alcohol above
threshold, the mean level of consumption was nine
drinks per day (mean = 8.78, SD = 5.36, n = 52). For
cannabis users, mean usage was 14 cannabis use occa-
sions (joints/bongs) per day (mean = 14.06, SD = 16.19,
n = 69) in the month prior to assessment. Participants
reported an average hazardous use index of 40.34
(SD = 18.21, range 4–124) for the previous month.
Sixty-seven participants (69%) completed all three
follow-up assessments, with comparable rates across the
conditions (BI = 70.0%, therapist = 65.7%, com-
puter = 71.9%). There were no significant differences
between those who completed all follow-up assessments
and those who did not with respect to age, gender, treat-
ment condition, alcohol/cannabis use threshold status or
hazardous use index scores. There were also no differen-
tial age or gender effects by treatment group. Participants
in the therapist-delivered group attended an average of
nine treatment sessions including the BI (mean = 8.71,
SD = 2.74, range 1–10 sessions) with those allocated to
the computer-delivery condition attending around eight
sessions (mean = 7.61, SD = 2.87, range = 2–10 ses-
sions; F(1,44) = 1.73, P = 0.20). Within these treatment
subgroups, 78% (n = 18) therapist-delivered SHADE
therapy participants attended all their allocated sessions,
compared with 52% (n = 12) within the computer-
delivered SHADE therapy condition. This difference was
not statistically significant (c21 = 2.396, P = 0.122).
Changes in symptoms over time
Table 1 displays the change in symptoms reported by the
67 participants who completed all baseline and follow-up
assessments.
Depression
As shown in Table 1, there was a significant reduction in
BDI-II scores over time across treatment groups (linear
trend, P < 0.001) which was more pronounced between
baseline and 3 months (quadratic trend, P < 0.001), with
some deterioration between 3 and 6 months, followed by
stabilization to 12 months (cubic trend, P < 0.001). The
therapist-delivered condition showed the most marked
reduction from baseline to 3 months but also the stron-

Table 1 Changes in symptoms by treatment group for participants who completed all assessment phases.

Assessment occasion

Outcome Baseline 3 months 6 months 12 months

Significant Significant group
Therapy condition Mean SD Mean SD Mean SD Mean SD time effects ¥ time interactions

Depressiona (n = 67)
BI 32.86 9.59 22.95 10.46 28.29 13.19 24.76 12.55 L: F(1,64) = 41.17** Q ¥ BI versus T:
F(1,64) = 13.73**
Therapist (T) 34.91 9.7 13.04 10.51 15.46 11.11 20.35 14.49 Q: F(1,64) = 38.72** Q ¥ T versus CM:
F(1,64) = 11.58**
Computer (CM) 28.57 9.89 17.09 12.14 16.65 10.63 13.65 9.55 C: F(1,64) = 30.18**
Alcoholb (n = 41)
BI 8.18 5.17 4.79 4.95 6.41 5.91 4.03 3.22 L: F(1,38) = 14.56** NS
Therapist (T) 9.6 5.45 3.58 4.6 3.62 5.31 2.49 3.47 C: F(1,38) = 9.18*
Computer (CM) 7.34 4.48 3.81 4.92 6.39 9.57 4.13 5.78
Cannabisb (n = 43)
BI 9.22 8.57 7.24 7.77 8 9.7 8.61 10.16 L: F(1,40) = 15.24** L ¥ BI versus T+CM:
F(1,40) = 7.23*
Therapist (T) 15.03 13.87 8.9 11.25 7.1 9.51 5.72 6.22
Computer (CM) 11.94 9.14 5.77 6.56 4.97 6.93 3.34 5.52
Hazardous usec (n = 67)
BI 39.67 15.4 31.11 13.54 38.78 17.14 34.11 16.01 L: F(1,64) = 21.36** L ¥ BI versus T+CM:
F(1,64) = 7.35*
Therapist (T) 43.11 17.04 39.84 50.27 27 22.8 24.21 18.71
Computer (CM) 42.1 20.17 23.43 16.92 25.76 14.58 21.05 11.75

Time effects: linear (L), quadratic (Q), cubic (C). *P < 0.01; **P < 0.001. aBeck Depression Inventory II scores. bUse occasions per day averaged over past
month (Opiate Treatment Index) for participants using alcohol or cannabis at harmful levels upon entry to the study. cUse of all substances, including
alcohol over past month (possible range 0–280-day equivalents). BI: brief intervention.

© 2009 The Authors. Journal compilation © 2009 Society for the Study of Addiction Addiction, 104, 378–388
 Computerized psychological treatment for comorbidity
gest level of relapse between 3 and 12 months (interac-
tions between quadratic trend and therapist versus brief
and computer conditions, P < 0.001).
Alcohol consumption
There was a significant overall reduction across treat-
ment groups in alcohol consumption among the 41 par-
ticipants who met baseline threshold for problematic
alcohol use (linear trend, P < 0.001; see Table 1). There
was some increase in drinking between 3 and 6 months,
followed by a reduction in drinking (to a level at least
equal to that achieved at 3 months) by the 12-month
follow-up point (cubic trend, P < 0.01). There were no
significant differences between intervention conditions
(i.e. no group or interaction effects).
Cannabis use
There was a significant overall reduction in cannabis use
over 12 months (linear trend, P < 0.001) among the 43
participants who met baseline threshold criteria for can-
nabis use. However, participants in the SHADE intensive
therapy conditions reported a significantly greater reduc-
tion from baseline to 12 months compared to the BI con-
dition (interaction with linear trend, P < 0.001; see
Table 1).
Hazardous AOD use days
There was a significant overall reduction in hazardous
use days over the 12-month follow-up period (linear
trend, P < 0.001). Participants in the SHADE intensive
therapy conditions reported a significantly greater reduc-
tion between baseline and 12-month follow-up compared
to the BI condition (47% reduction versus 14%, interac-
tion with linear trend, P < 0.01; see Table 1).
ITT analyses
Of the 97 people recruited, 82 (85%), 79 (81%) and 74
(76%) completed the 3-, 6- and 12-month follow-up
assessments, respectively. In the last observation carried
forward analyses, approximately 19.2% of follow-up
assessments were substituted from previous assessment
phases. The results based on all 97 participants were con-
sistent with the principal analyses, except for cannabis
use. Participants in the computer-delivered condition
reported a significantly greater reduction from baseline to
12 months compared to the BI condition (interaction
with linear trend; F(1,66) = 7.13, P = 0.01). The
computer-delivered participants reported a threefold
reduction in their level of use (baseline, 16.90; 12
months, 4.75), while the BI condition reported no
© 2009 The Authors. Journal compilation © 2009 Society for the Study of Addiction
383
demonstrable change (baseline, 10.81; 12 months,
10.05).
Primary outcome measures at the 12-month
critical time-point
The left-hand columns of Table 2 report raw and stan-
dardized (ES) change scores from baseline to 12 months.
ITT and non-ITT analyses of categorical outcome mea-
sures at 12 months, according to treatment condition,
are displayed in the right-hand columns of Table 2. In
each case, ITT analyses confirmed the non-ITT results,
which are described in detail below. Differences in treat-
ment conditions of less than 0.25 SD (ES) units were not
considered to be clinically important.
The standardized change scores in Table 2 were
obtained by dividing the raw change scores by the rel-
evant pooled SD. For example, relative to a reference
pooled SD of 12.98 units on the BDI-II, computer- (ES
0.98) and therapist-delivered (ES 1.07) treatments were
associated with relatively large but nevertheless compa-
rable standardized reductions at 12 months. There was
some evidence of greater improvement in depression for
SHADE intensive versus BI (ES difference 0.31), therapist
versus BI (ES difference 0.36) and computer versus BI (ES
difference 0.27) conditions.
As indicated in Table 2, there were sizeable ES differ-
ences between BI and therapist-delivered treatment for
cannabis use (ES difference 0.76), and between BI and
computer-delivered treatment (ES difference 1.11, refer-
ence pooled SD = 12.21). Findings in the same direction
were found for hazardous use days, with ES differences of
0.48 between BI and therapist interventions and 0.78
differences between BI and computer interventions
(reference pooled SD = 22.29). In contrast, for alcohol
consumption, the difference between therapist- versus
computer-delivered interventions at 12-month follow-up
was 0.36, in favour of therapist-delivered treatment (ref-
erence pooled SD = 5.34). There were no differences
between BI and computer interventions for alcohol con-
sumption (ES difference 0.01), with larger differences
between BI and therapist-delivered interventions (ES dif-
ference 0.37).
As shown in the right-hand columns of Table 2,
there were no significant differences between SHADE
intensive conditions in the percentage of subjects report-
ing improvement, as defined by scores <17 on the BDI-II,
or at least a 50% reduction in alcohol or cannabis use.
There was only one statistically significant difference in
the categorical outcome analyses. Participants in the
computer-delivered condition reported significantly
better AOD outcomes compared to the BI condition,
with computer-delivered subjects five times more likely
[odds ratio (OR) = 5.24, or OR = 5.00 in the ITT
Addiction, 104, 378–388
384 Frances J. Kay-Lambkin et al.

Table 2 Improvement in key outcomes at 12-month follow-up assessment.

Outcome Raw Standardized % Improved

Therapy condition change (ITT) change (ITT) (ITT)a OR (99% CI)b (ITT)

Depressionc (n = 67)
BI 9.26 (7.73) 0.71 (0.60) 30.4 (26.7) 1
Therapist (T) 13.88 (9.91) 1.07 (0.76) 50.0 (37.1) 2.29 (0.48, 11.00)
[1.63 (0.40, 6.6)]
Computer (CM) 12.74 (11.44) 0.98 (0.88) 63.0 (56.3) 3.89 (0.82, 18.39)
[3.54 (0.87, 14.41)]
T+CM 13.27 (10.64) 1.02 (0.82) 56.9 (46.3) 3.01 (0.76, 11.93)
[2.37 (0.69, 8.16)]
Alcohold (n = 41)
BI 4.56 (4.00) 0.85 (0.75) 53.8 (50.0) 1
Therapist (T) 6.52 (5.72) 1.22 (1.07) 82.4 (75.0) 4.00 (0.45, 35.28)
[3.00 (0.47, 19.11)]
Computer (CM) 4.57 (4.46) 0.86 (0.84) 73.3 (75.0) 2.36 (0.30, 18.82)
[3.00 (0.42, 21.46)]
T+CM 5.60 (5.16) 1.05 (0.97) 78.1 (75.0) 3.06 (0.50, 18.66)
[3.00 (0.59, 15.24)]
Cannabisd (n = 43)
BI 0.82 (0.76) 0.07 (0.06) 44.4 (34.8) 1
Therapist (T) 10.13 (6.22) 0.83 (0.51) 61.5 (45.5) 2.00 (0.30, 13.51)
[1.56 (0.32, 7.57)]
Computer (CM) 14.43 (12.15) 1.18 (1.00) 78.9 (70.8) 4.69 (0.70, 31.21)
[4.55 (0.91, 22.91)]
T+CM 12.68 (9.31) 1.04 (0.76) 71.9 (58.7) 3.19 (0.65, 15.62)
[2.66 (0.68, 10.45)]
Hazardous usee (n = 67)
BI 5.61 (2.50) 0.25 (0.11) 21.7 (16.7) 1
Therapist (T) 16.17 (11.09) 0.73 (0.50) 37.5 (28.6) 2.16 (0.40, 11.77)
[2.00 (0.41, 9.79)]
Computer (CM) 23.04 (20.34) 1.03 (0.91) 59.3 (50.0) 5.24 (1.01, 27.19)*
[5.00 (1.06, 23.70)]*
T+CM 19.80 (15.51) 0.89 (0.70) 49.0 (38.8) 3.46 (0.78, 15.34)
[3.17 (0.77, 13.09)]

*P < 0.01. a% Improved: % of sample reporting scores <17 BDI, >49% improvement in alcohol, cannabis or hazardous use days. bOdds ratios (OR) and
associated 99% confidence intervals (CI), with the brief intervention (BI) group as the reference point (OR = 1.00): intention-to-treat (ITT), missing
data were brought forward from the previous assessment phase for which there was a score. cBeck Depression Inventory II scores. dUse occasions per day
averaged over past month (Opiate Treatment Index) for participants using alcohol or cannabis at harmful levels upon entry to the study. eUse of all
substances, including alcohol over past month (possible range 0–280-day equivalents).

analyses] to report a reduction of at least 50% in haz-
ardous use days.
Treatment fidelity
BI sessions (session 1) lasted an average of 66.56
minutes (SD = 15.33, range = 45–95), therapist-
delivered sessions (sessions 2–9) lasted an average of
59.58 minutes (SD = 16.19, range = 31–94) and
check-in sessions for the computer condition lasted an
average of 12.31 minutes (SD = 8.17, range = 3–37).
Across treatment conditions, participants allocated to BI
alone received 66.56 minutes of therapist time,
therapist-delivered participants received 602.78
minutes over 10 sessions and computer-based
participants received 177.35 therapist minutes over 10
sessions (including the BI as session 1). Of the 73 tapes
rated for treatment fidelity, 47 (64%) were assessed as
being fully adherent to the agenda for that particular
session. Fourteen BI sessions (78%), nine therapist-
delivered sessions (35%) and 24 check-in sessions (83%)
were rated as adherent, which was statistically signifi-
cant (c22 = 15.80, P = 0.00). The majority of reasons for
non-adherence were not considered to be significantly
deviant from the treatment manual for the study (e.g.
MI not covered/required, agenda abandoned due to sui-
cidal crisis, no explicit review of week, [37]). However,
in six cases in both the BI and therapist-delivered
sessions, and in three computer check-ins, treat-
ment exceeded the maximum allowable time (e.g.

© 2009 The Authors. Journal compilation © 2009 Society for the Study of Addiction Addiction, 104, 378–388
 Computerized psychological treatment for comorbidity
approximately 60 minutes for therapy sessions and 10
minutes for check-ins).
DISCUSSION
Among this sample of people with comorbid major
depression and alcohol and/or cannabis use problems,
differences in treatment efficacy were evident across
symptom domains and classes of substance.
Efficacy of computer-based interventions
Regardless of treatment allocation, participants reported
significant reductions in depression, alcohol and can-
nabis use. These results were confirmed by ITT analysis.
Rates of participation and treatment retention were
equivalent between the treatment modalities, potentially
indicating support for the acceptability of computer-
based treatments among this population.
Levels of depression
Computer- and therapist-delivered treatments could be
regarded as producing similar benefits for people with
comorbid depression and AOD use problems, with
similar depression profiles at 12 months. Both therapist-
and computer-based interventions were associated with
differential effect sizes larger than a quarter of a SD
relative to the BI (on non-ITT analysis), a threshold
regarded as evidence of non-equivalence (or potential
differential clinical utility) for the purposes of this
study.
Overall, the sample reported significant reductions in
BDI-II scores over the follow-up period. However, the
pattern of change in depression suggests a more immedi-
ate response to treatment among those in the therapist-
delivered condition, although relapse rates were highest
among this group between 3- and 12-month follow-ups.
Computer-based clients took longer to match the reduc-
tions in depression reported by the therapist condition;
however, at 12 months this group reported lower levels of
depression, scoring in the minimal range for depressive
symptoms, although their baseline scores were also
lower. It is unclear why the therapist-delivered group did
not maintain the significant changes at 12 months rela-
tive to the other occasions. In the short term, the direc-
tion of a ‘live’ therapist may have resulted in participants
grasping more quickly the integrated concepts of SHADE
intensive therapy, and closer monitoring may have
encouraged them to practice these strategies sooner than
did participants in the computer condition. However, the
self-help nature of the computer-delivered treatment may
have benefited participants over the longer term. Further
© 2009 The Authors. Journal compilation © 2009 Society for the Study of Addiction
385
research on maintaining the early improvements in
depression seen in the therapist-delivered intervention is
worthy of investigation.
AOD outcomes
There was a different pattern of results for alcohol versus
cannabis use and hazardous use days. The BI and
computer-based interventions were associated with mod-
erate to large effect sizes for alcohol consumption at 12
months, with the therapist-delivered intervention pro-
ducing the largest effect. The converse was true for can-
nabis use and hazardous use days (largely reflecting the
cannabis use results). BI was less effective in reducing
cannabis use and hazardous use days at 12 months com-
pared to SHADE intensive interventions and, while both
types of SHADE intensive therapy were associated with at
least moderate effect sizes, computer-delivered therapy
appeared to be associated with a larger effect size than
that delivered by a therapist.
The alcohol outcomes observed in this study approxi-
mate closely those reported in the Project MATCH study,
comparing alcohol use outcomes for people engaged in a
brief motivational intervention versus CBT and a 12-Step
facilitation programme [38]. That is, at the 12-month
follow-up assessment, Project MATCH participants
assigned to receive the brief motivational intervention
reported equivalent alcohol use outcomes relative to their
counterparts who received more intensive treatment.
Consistent with the results of this study, Project MATCH
participants with higher levels of psychiatric severity did
not respond better to the more intensive CBT or 12-Step
treatment relative to the brief motivational intervention.
Taken together, these findings suggest a potential benefit
of BIs, in addition to assessment, as a first step in
treatment for people with depression and alcohol use
problems, regardless of severity. The next step for non-
responders would seem to be therapist-delivered interven-
tions, given the indications of superior therapist- versus
computer-delivered intervention outcomes for alcohol.
These steps could, potentially, be applied by clinicians
working in a variety of different treatment settings,
including mental health, drug and alcohol and primary
care.
For cannabis use, intensive, preferably computer-
based, interventions appear most efficacious for people
with coexisting depression and cannabis use problems.
However, despite important reductions in daily levels of
cannabis use, 12-month levels remained high across the
treatment conditions, with participants continuing to use
between three and nine times daily. This is in contrast to
levels of alcohol use, which dropped to within recom-
mended safe drinking guidelines by the 12-month assess-
ment. It is unclear why this occurred. A similar result was
Addiction, 104, 378–388
386 Frances J. Kay-Lambkin et al.
reported by Baker et al. [37], where cannabis use at 12
months remained above the hazardous threshold of
once-weekly use (range 4.12–8.53 use occasions per day)
and alcohol use dropped to within recommended safe
drinking levels. The difficulty in shifting cannabis use to
low levels and the unique combination of cannabis use
and depression comorbidity warrants further attention.
Limitations
In general, the study results were confirmed by ITT analy-
sis, although it is acknowledged that an assumption
inherent in the last observation carried forward analysis
is that participants remain stable over the assessment
time-points for which data are being carried forward. This
may not have been the case, particularly if dropout was
associated with an exacerbation in symptoms, making
attendance at scheduled appointments to complete the
treatment programmes difficult. The small sample size
reduced the power of the current study to detect differ-
ences between treatment groups and replication is
required to further explore these observations. It is also
possible that the patterns of change associated with the
study are due to a high level of motivation for change
among the self-referred study participants and, therefore,
may not represent treatment attendance and outcomes
accurately among a less-motivated sample. However, as
people with comorbid depression and AOD use problems
do not typically access treatment within mental health or
AOD settings, it may be that these participants represent,
at least partly, the group of people with this comorbidity
within the community.
Primacy of depression or alcohol/cannabis use was
not established for this study, although all participants
met criteria for life-time major depressive disorder. Not
distinguishing between primary/secondary diagnoses
may cause problems for proponents of some models of
comorbidity, which suggest that treatment of the primary
condition will result in resolution of the second condition
[39]. However, considering the difficulties in determining
primacy for long-term conditions and the outcome pro-
files for depression and AOD use reported above, there
appears to be little justification for withholding treatment
for one of these conditions.
Future replications of this study would also need to
consider a control group matched for therapist contact,
rather than offering a single 60-minute intervention.
This may help to determine the extent to which the pat-
terns reported above were attributable to the treatment
strategies and orientation, as opposed to increased thera-
pist contact. In addition, the computer-based interven-
tion was delivered with weekly therapist contact, making
it difficult to determine the unique contribution of the
computerized treatment on the results. Previous research
© 2009 The Authors. Journal compilation © 2009 Society for the Study of Addiction
on depressive and anxiety disorders has indicated that the
addition of minimal therapist contact within a computer-
based intervention, such as that utilized in this study, can
be associated with an increased effect size relative to
computer-based intervention without such contact [16].
Future replications might consider including a computer-
only condition. Notwithstanding these limitations, the
results show promise for the benefits of integrated
psychological treatment for depression and alcohol or
cannabis use comorbidity, delivered via therapist or com-
puter, and are worthy of further exploration.
Summary
Little previous research has been conducted on the ben-
efits of integrated psychological treatment, targeting
both mental disorders and substance use problems,
among people with comorbidity, especially depression
and comorbid substance use problems. The main findings
of this study were: (i) depression responded better to
intensive MI/CBT treatment compared to BI alone, with
therapist-delivered treatment demonstrating a strong
short-term beneficial effect for depression, which was
matched by the computer-delivered treatment partici-
pants by the 12-month follow-up assessment; (ii) prob-
lematic alcohol consumption responded well to a BI and
even better to the intensive MI/CBT intervention, particu-
larly when therapist-delivered; and (iii) that intensive
MI/CBT was significantly better than BI alone in reducing
cannabis use and hazardous use of substances, with
computer-based therapy showing the largest treatment
effect. Computer-delivered treatment supplemented by
brief therapist support (average 12 minutes per session),
however, in comparison to therapist-delivered treatment
(average 60 minutes per session), saved approximately
79% of therapist time over the course of the trial, and
produced similar outcomes in both depressive and sub-
stance use domains at 12-month follow-up.
Declarations of interest
None.
Acknowledgements
The authors wish to acknowledge the involvement of the
study participants, without whom this research would
not be possible. The study was funded in part by the
Alcohol-related Medical Research Scheme (Australian
Brewer’s Foundation) and a bequest from Ms Jennie
Thomas on behalf of her late husband Philip Emelyn
Thomas via the University of Newcastle, Australia. In
addition, a National Health and Medical Research
Council (NHMRC) public health postgraduate scholar-
ship supported the primary author. The research team
Addiction, 104, 378–388
 Computerized psychological treatment for comorbidity
remained independent from the funding bodies. This
study was carried out in accordance with the National
Health and Medical Research Council of Australia’s
Statement of Ethical Conduct of Research among Human
Participants. Ethics approval was gained from the rel-
evant Human Research Ethics Committees (HAREC
Approval no. 02/03/13/3.16, HREC Approval no. H 307
0502).
References
1. Andrews G., Henderson S., Hall W. Prevalence, comorbidity,
disability and service utilisation: overview of the Australian
National Mental Health Survey. Br J Psychiatry 2001; 178:
145–53.
2. Andrews G., Issakidis C., Slade T. How common is comor-
bidity? In: Teesson M., Proudfoot H., editors. Comorbid
Mental Disorders and Substance Use Disordesr: Epidemiology,
Prevention and Treatment. Canberra: Commonwealth of
Australia (publication approval no. 3415); 2003; p. 26–41.
3. Clarkin J. F., Kendall P. Comorbidity and treatment plan-
ning: summary and future directions. J Consult Clin Psychol
1992; 60: 904–8.
4. Tate D. F., Zabinski M. F. Computer and internet applications
for psychological treatment: update for clinicians. J Clin
Psychol 2004; 60: 209–20.
5. Proudfoot H., Ryden C., Everitt B., Shapiro D. A., Goldberg
D., Mann A. et al. Clinical efficacy of computerised
cognitive-behavioural therapy for anxiety and depression in
primary care: randomised controlled trial. Br J Psychiatry
2004; 185: 46–54.
6. Christensen H., Griffiths K. M., Jorm A. F. Delivering inter-
ventions for depression by using the internet: randomised
controlled trial. BMJ 2004; 328: 265–9.
7. Selmi P. A., Klein M. H., Greist J. H., Sorrell S. P., Erdman H.
P. Computer-administered cognitive-behavioral therapy for
depression. Am J Psychiatry 1990; 147: 51–6.
8. White J., Jones R., McGarry E. Cognitive behaviour com-
puter therapy for the anxiety disorders: a pilot study. J Ment
Health 2000; 9: 505–16.
9. Carlbring P., Westling B., Ljungstrand P., Ekselius H. P.,
Andersson G. Treatment of panic disorder via the Internet:
a randomized trial of a self-help program. Behav Ther 2001;
32: 751–64.
10. Shiffman S., Paty J. A., Rohay J. M., Di Marino M. E., Gitchell
J. G. The efficacy of computer-tailored smoking cessation
material as a supplement to nicotine patch therapy. Drug
Alcohol Depend 2001; 64: 35–46.
11. Hester R. K., Delaney H. D. Behavioral self-control program
for Windows: results of a controlled clinical trial. J Consult
Clin Psychol 1997; 65: 686–93.
12. Linke S., Brown A., Wallace P. Down your drink: a web-
based intervention for people with excessive alcohol con-
sumption. Alcohol Alcohol 2004; 39: 29–32.
13. Kerwin M-LE. Evaluation of a computer-based instructional
package about eating disorders. Comput Hum Behav 2006;
22: 1059–66.
14. Copeland J., Martin G. Web-based interventions for sub-
stance use disorder: a qualitative review. J Subst Abuse Treat
2004; 26: 109–16.
15. Kypri K. J., Saunders J. B., Williams S. M., McGee R. O.,
Langley J. D., Cashell-Smith M. L. et al. Web-based screening
© 2009 The Authors. Journal compilation © 2009 Society for the Study of Addiction
387
and brief intervention for hazardous drinking: a double-
blind randomized controlled trial. Addiction 2004; 99:
1410–7.
16. Spek V., Cuijpers P., Nyklicek I., Riper H., Keyzer J., Pop V.
Internet-based cognitive behaviour therapy for symptoms of
depression and anxiety: a meta-analysis. Psychol Med 2007;
37: 319–28.
17. Carlbring P., Nilsson-Ihrfelt E., Waara J., Kollenstam C.,
Buhrman M., Kaldo V. et al. Treatment of panic disorder:
live therapy versus self-help via the Internet. Behav Res
Therapy 2005; 43: 1321–33.
18. National Institute for Clinical Excellence (NICE). Guidance on
the Use of Computerised Cognitive Behavioural Therapy for
Anxiety and Depression. Technology Appraisal no. 51.
London: NICE; 2002.
19. Hunter Area Health Service (HAHS). Hunter Health Annual
Report: 2003–2004. Newcastle, NSW: HAHS; 2004, p. 363.
20. Beck A. T., Steer R. A., Brown G. K. The Beck Depression
Inventory, 2nd edn. Manual. San Antonio: The Psychologi-
cal Corporation; 1996.
21. First M. B., Spitzer R. L., Gibbon M., Williams J. B. W. Struc-
tured Clinical Interview for DSM-IV-TR Axis I Disorders,
Research Version, Patient edition. New York: Biometrics
Research, New York State Psychiatric Institute; 2001.
22. Shand F., Gates J., Fawcett J., Mattick R. The Treatment of
Alcohol Problems: A Review of the Evidence. Canberra: Aus-
tralian Commonwealth Department of Health and Ageing;
2003.
23. NSWHealth. The management of people with a co-existing
mental health and substance use disorder: discussion paper.
Report no.: State Health Publication no.: (CMH) 000050.
NSW: New South Wales Health Department Dual Diagnosis
Project; 2000.
24. Moggi F., Brodbeck J., Koltzsch K., Hirsbrunner H-P. One-
year follow-up of dual diagnosis patients attending a four-
month integrated inpatient treatment. Eur Addict Res 2002;
8: 30–7.
25. Drake R., Mueser K., Brunette M. F., McHugo G. J. A review
of treatment for people with severe mental illness and
co-occurring susbtance use disorders. Psychiatr Rehabil J
2004; 27: 360–74.
26. Carroll KM. Behavioral therapies for co-occurring sub-
stance use and mood disorders. Biol Psychiatry 2004; 56:
778–84.
27. Mueser K. T., Noordsy D. L., Drake R. E., Fox L. Integrated
Treatments for Dual Disorders: A Guide to Effective Practice.
New York: Guilford Press; 2003.
28. Drake R., Mueser K. Psychosocial approaches to dual diag-
nosis. Schizophr Bull 2000; 26: 105–18.
29. Miller W. R., Rollnick S. Motivational Interviewing: Preparing
People to Change Addictive Behaviour. New York: Guilford
Press; 1991.
30. Beck A. T., Wright F. D., Newman C. F., Liese BS. Cognitive
Therapy of Substance Abuse. New York: Guilford Press; 1993.
31. Beck A. T., Rush A. J., Shaw B. F., Emery G. Cognitive Therapy
of Depression. New York: Guilford Press; 1979.
32. Segal Z. V., Williams J. M. G., Teasdale JD. Mindfulness-Based
Cognitive Therapy for Depression: A New Approach to Prevent-
ing Relapse. New York: Guilford Press; 2002.
33. Persons J. B., Davidson J., Tompkins MA. Essential Compo-
nents of Cognitive-Behaviour Therapy for Depression. Wash-
ington, DC: American Psychological Association; 2001.
34. Graham H. L., Birchwood M. J., Mueser K. T., Orford J.,
McGovern D., Atkinson E. et al. Cognitive-Behavioural
Addiction, 104, 378–388
388 Frances J. Kay-Lambkin et al.
Integrated Treatment (C-BIT): A Treatment Manual for Sub-
stance Misuse in People with Severe Mental Health Problems.
West Sussex: John Wiley & Sons Ltd; 2004.
35. Tarrier N., Wells A. Treating Complex Cases. London: Wiley &
Sons; 1998.
36. Darke S., Ward J., Hall W., Heather N., Wodak A. The Opiate
Treatment Index (OTI) Manual. Sydney: National Drug and
Alcohol Research Centre; 1991.
37. Baker A., Bucci S., Lewin T. J., Kay-Lambkin F. J., Carr V. J.,
Constable P. Randomised controlled trial of cognitive
© 2009 The Authors. Journal compilation © 2009 Society for the Study of Addiction
behavior therapy for substance use disorders among
people with a psychotic illness. Br J Psychiatry 2006; 188:
439–48.
38. Project MATCH Research Group (PMRG). Matching alco-
holism treatments to client heterogeneity: Project MATCH
posttreatment drinking outcomes. J Stud Alcohol 1997; 58:
7–29.
39. Mueser K. T., Drake R. E., Wallach M. A. Dual diagnosis: a
review of etiological theories. Addict Behav 1998; 23: 717–
34.
Addiction, 104, 378–388