You are on page 1of 8

Behaviour Research and Therapy 51 (2013) 344e351

Contents lists available at SciVerse ScienceDirect

Behaviour Research and Therapy

journal homepage:

Internet-based guided self-help for university students with anxiety,

depression and stress: A randomized controlled clinical trial
Victor Day a, *, Patrick J. McGrath b, c, Magdalena Wojtowicz c

Counselling Centre, Students Union Building, Dalhousie University, Halifax, NS B3H 4R2, Canada
IWK Health Centre and Dalhousie University, Canada
Department of Psychology and Neuroscience, Dalhousie University, Canada

a r t i c l e i n f o

a b s t r a c t

Article history:
Received 19 November 2012
Received in revised form
1 March 2013
Accepted 11 March 2013

Anxiety, depression and stress, often co-occurring, are the psychological problems for which university
students most often seek help. Moreover there are many distressed students who cannot, or choose not
to, access professional help. The present study evaluated the efcacy of an internet-based guided selfhelp program for moderate anxiety, depression and stress. The program was based on standard cognitive behavior therapy principles and included 5 core modules, some of which involved options for
focusing on anxiety and/or depression and/or stress. Trained student coaches provided encouragement
and advice about using the program via e-mail or brief weekly phone calls. Sixty-six distressed university
students were randomly assigned to either Immediate Access or a 6-week Delayed Access condition.
Sixty-one percent of Immediate Access participants completed all 5 core modules, and 80% of all participants completed the second assessment. On the Depression, Anxiety and Stress Scales-21, Immediate
Access participants reported signicantly greater reductions in depression (h2p . 07), anxiety (h2p . 08)
and stress (h2p . 12) in comparison to participants waiting to do the program, and these improvements
were maintained at a six month follow-up. The results suggest that the provision of individuallyadaptable, internet-based, self-help programs can reduce psychological distress in university students.
2013 Elsevier Ltd. All rights reserved.

Guided internet-based treatment
Randomized controlled trial (RCT)
University students

Anxiety and depression, experienced to various degrees and

often in combination, are the most prevalent mental health problems (Kessler, Demler, & Frank, 2005). This general phenomenon is
also true for the university student population (Adalf, Gliksman,
Demers, & Newton-Taylor, 2001). For example Price, McLeod,
Gleich, and Hand (2006) found that amongst students at a Canadian university the one year prevalence rate for Major Depressive
Disorder was 7% for men and 14% for women, and for an Anxiety
Disorder was 13% for men and 19% for women, and in all instances
the number of students with a level of subclinical depression or
anxiety that potentially put them at risk for developing a clinical
disorder was at least twice the number of those who met the
criteria for a clinical disorder. Within that same sample, 6% of male
students and 11% of female students had both a Depressive Disorder and an Anxiety Disorder. Anxiety and depression are the most
common problems for which students seek psychological help
(Krumrie, Newton, & Kim, 2010). In the 2010 American College
Health Survey, on a survey question asking if and how often they

* Corresponding author. Tel.: 1 902 494 2081.

E-mail address: (V. Day).
0005-7967/$ e see front matter 2013 Elsevier Ltd. All rights reserved.

experienced overwhelming anxiety, 48% of college and university

respondents felt overwhelming anxiety at least once in the preceding 12 months; and on an item asking if and how often they felt
so depressed it was difcult to function, 31% reported feeling so
depressed it was difcult to function at least once in the preceding
year (ACHA, 2010).
There are effective, empirically supported psychological treatments for anxiety and depression, most of them involving cognitive-behavioral therapy (CBT) (Chambless & Ollendick, 2001).
However, most distressed people, including most students, do not
seek psychological treatment (Eisenberg, Goldberstein, & Gallant,
2007; Harrar, Affspring, & Long 2010; Wang, Lang, & Olfsen,
2005). There are various possible reasons for this, including concerns about stigma, embarrassment and privacy, skepticism about
treatment, cost, convenience, etc. (Vogel, Wester, & Larson, 2007).
However, simply entreating students to seek therapy is not a realistic solution, since there are often insufcient resources to serve
the students who do seek psychological help. North American
university counseling services are often unable to meet current
demands for services (Eisenberg et al., 2007; Kitzrow, 2003).
Alternate solutions must be found, if more students are to be helped. Since the essence of CBT involves learning new ways of

V. Day et al. / Behaviour Research and Therapy 51 (2013) 344e351

thinking and acting, it is possible for such learning to take place

outside of individual or group therapy. This has been demonstrated
in the past through the biblio-therapeutic use of books and manuals (Scogin, Bynum, & Stephens, 1990).
The current generation of university students reports a high
level of comfort with computers and a very high level of use of the
internet for many activities, including gathering information,
business and socializing (Caruso & Salaway, 2008). Kittinger,
Correia, and Irons (2012) found that at a large American university, 97% of students had access to the internet at their primary
residence, and they spend an average of 14 h per week on-line.
Computer programs have the potential to be more interactive and
individualized, and therefore more engaging, than written materials. However, computer-based programs that require going to a
clinic location to use are not favored by university students
(Mitchell & Gordon, 2007). Therefore program delivery via the
internet, which is more convenient, exible and private for the user,
appears to be the most promising method of delivery.
Computer programs available via the internet have been
developed and found to be helpful for various psychological
problems, including anxiety and depression (Anderson, 2009; Spek
et al., 2007). Internet based programs for social phobia (Tillfors
et al., 2008), bulimia (Sanchez-Ortz et al., 2011) and perfectionism (Radhu, Daskalkis, Arpin-Cribie, Irvine, & Ritvo 2012) have
been used effectively with university students. Meta-analyses of
studies of internet based programs have concluded that programs
which included some degree of personal guidance or at least regular e-mail or other contact with program staff produced better
results than programs which were entirely self-help (Andersson &
Cuijpers, 2009; Spek et al., 2007). However, Titov et al. (2009, 2010)
as well as Robinson et al. (2010) found that it is not necessary for
the person providing such guidance to be a professionally trained
therapist. So although guided self-help appears to be more effective
than pure self-help, it may be more cost effective to have the
guidance be provided by paraprofessional level personnel.
Most studies of the effectiveness of guided internet-based CBT
for psychological distress have been on specic anxiety disorders
(e.g. social anxiety, panic disorder, health anxiety) or depression
(Spek et al., 2007). Of course this permits for more homogeneous
research participant groups. However, specic anxiety problems,
and even anxiety and depression, are highly inter-correlated, and
many people need help for combinations of these problems (Brown
& Barlow, 1992). Moreover the CBT solutions for these problems are
very similar, and normal clinical practice outside of research trials
often includes help for both anxiety and depression to various
degrees. It is arguably more efcient and convenient for users to
have one self-help program which addresses both anxiety and
depression. Titov et al. (2011) recently reported on a controlled
clinical trial nding that a transdiagnostic on-line CBT program
for people with depression, or anxiety disorders, or both depression
and anxiety disorders, signicantly reduced both depression and
anxiety. It is notable that although their participants just had to
have an anxiety disorder or depression, most had both.
Some components of CBT, such as changing dysfunctional
thinking, are generically applicable to both anxiety and depression
(Beck, Benedict, & Winkler 2003); while other components are
relevant mostly for specic problems, such as behavioral activation
for depressed inactivity (Martell, Dimidjian, & Herman-Dunn,
2010). Broadening the potential applicability of a self-help program, in terms of problems addressed, does imply that users will be
able to partially individualize the program to t their specic needs
and goals. With perhaps the exception of manualized treatments
for specic problems within clinical research trials, individual inperson CBT for anxiety and depression often includes various
negotiated choices about specic foci and methods, even though


there is a common core of changing distressing thoughts and

related problematic behaviors that are causing the problems to
persist (Beck, 2011). Replicating this type of desirable exibility
within guided self-help programs requires providing users with
some choices about specic foci and methods. Recently Carlbring
et al. (2011) reported on the signicant effectiveness of a guided
self-help program applicable a variety of anxiety problems in which
participants also received an in-person interview and individualized advice about which modules to use. Johansson et al. (2012)
compared the effectiveness of a standardized internet based program for depression with a similar program which also included
additional material on strategies for anxiety and stress which were
individually selected for the participant by therapists. They found
both programs were overall equally effective in comparison to a
control group, but the tailored treatment was more effective for a
subgroup of participants with more severe depression and more
co-morbid anxiety and other problems. Andersson, Estling,
Jakobsson, Cuijpers, and Carlbring (2011) have carried this progression a step further, and piloted a guided self-help program for
people with mixed anxiety disorders in which participants made
their own decisions about which modules to use based mainly on
descriptions provided within the program. Andersson et al. (2011)
recommended that the role of choice and tailoring should be
further explored in controlled trials and that patient choice could
be incorporated into Internet-delivered treatment packages.
This article reports the results of a randomized controlled clinical trial on the effectiveness of a guided self-help program for
moderate anxiety, depression and/or stress experienced by university students. The program is based on standard CBT principles
and includes ve core modules, some with options to learn basic
behavioral methods of graduated increases in activity to overcome
depression, and/or overcoming anxiety-related avoidance in hierarchical steps, and/or reducing problematic behaviors such as
overeating or drinking alcohol in reaction to stress; others with
generally applicable strategies for self-motivation for change such
as decisional balancing and specic goal-setting, and how to
change thoughts and underlying beliefs that cause anxiety and/or
depression. After completing the core modules participants also get
access to ve (for men) or six (for women) optional modules for
possibly related issues such as social relationships, sleep problems
and PMS. The usability of the content of the modules was initially
assessed and improved to a high level as reported by Currie,
McGrath, and Day (2010). Participants were given the opportunity to individualize the program by choosing to focus on streams
relevant for anxiety and/or depression, and/or stress. Participants
were also provided with program coaches (i.e. trained graduate or
undergraduate students) who provided periodic encouragement
and advice about using the program.
Our main hypotheses were that participants using the program
would improve with respect to anxiety, depression and stress
symptoms more so than students waiting to use the program, and
that these improvements would be maintained at a six-month
Participants and recruitment
Participants were recruited from Dalhousie University and the
Universities of Kings College and Nova Scotia College of Art and
Design, in Halifax, Canada. Recruitment for the study began in 2010.
Participants were recruited primarily via emails, advertisements in
a campus newspaper, and recruitment posters. Interested individuals contacted the primary program coach through email and
were provided with information regarding the study. The study


V. Day et al. / Behaviour Research and Therapy 51 (2013) 344e351

protocol was initially presented to participants on the websites

consent form page. Interested participants signed up for the program online and received a phone call from the primary program
coach who reviewed the consent form with the participant and
assessed their eligibility for the study. Participants who provided
verbal and written informed consented and met study criteria were
randomized into the study.
Participants were university students attending the aforementioned post-secondary institutes. Eligible participants met the
following criteria: (1) experiencing mild to moderate levels of
anxiety, depression or stress, (2) not experiencing suicidal thoughts,
(3) not experiencing symptoms of bipolar disorder, an eating disorder, an addiction, or psychosis, and (4) not receiving psychological
counseling nor recently started psychotropic medication.
140 participants inquired about the study, of those only 69
expressed interest in continuing with the study (i.e. responded to
contact, completed module 1, meet all inclusion/exclusion criteria),
and nally 66 were continued with the phone-consent procedure
and were randomized into the study (See CONSORT diagram Fig. 1).
Procedure and design
Once the primary program coach reviewed inclusion/exclusion
criteria, received informed consent, and initial online self-report
measures (i.e. scores of at least 10 for depression, at least 8 for
anxiety, or at least 15 for stress on the DASS-21), participants were
randomized into one of two conditions.
 Immediate-Access Group: Participants received immediate
access to the online program. Participants were given approximately 6 weeks to complete the program and were contacted

6 months after program completion to ll out follow-up

 Delayed-Access Group: Participants were asked to wait 6
weeks prior to commencing the program. After they completed
the wait-period they were provided with access to the program
and were contacted 6 months post-completion to ll out
follow-up measures.
All participants were assessed using a self-report questionnaire
at baseline (Assessment 1) and after approximately 6 weeks (i.e.
after the Immediate-Access group completed the program and the
Delay-Access group completed the wait-period; Assessment 2).
Participants in the Delayed-Access group who chose to complete
the program after the wait-period were assessed after program
completion (Assessment 3). All participants who completed the
program were also assessed 6 months after program completion (6
month follow-up).
Participants were not compensated for their participation and
the program was offered to participants free of charge. Ethical
approval for the study was received from the Dalhousie University
Research Ethics Board.
A computerized random permuted block sequence generator
called Random Allocation Software http://mahmoodsaghaei.tripod.
com/Softwares/randalloc.html was used to generate the allocation
sequence for the study. The allocation placement ratio was set at
50% to ensure equal allocation to the intervention and control
groups. The allocation placements were securely concealed in a
double-envelope system. Study investigators, including the

Inquired about study


Excluded (Decided to
seek personal counseling)
N= 3

Did not complete module 1

N= 61
Did not respond to contact
N= 10

Assessed for eligibility


Randomized (N=66)

Allocated to Immediate Group N=33

Received intervention N= 33(100%)
Did not receive intervention N=0 (0%)
Stopped between modules 1 to 3 N=11
Stopped between modules 4 to 5 N=2
Completed all modules N=20 (61%)

Allocated to Delayed Group N=33

Received intervention N= 33 (100%)
Did not receive intervention N=0 (0%)
Stopped between modules 1 to 3 N= 22
Stopped between modules 4 to 5 N=1
Completed all modules N= 10 (30.3%)

Analyzed N=33
Excluded from analyses N= 0
Completed 1st assessment N= 33 (100%)
Completed 2nd assessment N= 24
Completed 3rd assessment N= N/A
Completed 6 month follow-up N= 17

Analyzed N=33
Excluded from analyses N= 0
Completed 1st assessment N= 33 (100%)
Completed 2nd assessment N= 29
Completed 3rd assessment N= 14 (42.4%)
Completed 6 month follow-up N= 7

Fig. 1. Consort diagram.

V. Day et al. / Behaviour Research and Therapy 51 (2013) 344e351

primary program coach, remained blind to the randomization of

participants. The primary program coach opened the doubleenvelope only after participants provided informed consent over
the phone.
Primary and secondary outcome measures
The primary outcome measure was the Depression Anxiety
and Stress Scale-21 (DASS-21). The DASS-21 (Lovibond &
Lovibond, 1995) is an abbreviated version of the original 42item DASS. It is composed of three 7-item subscales measuring
symptoms of depression, anxiety, and stress. Participants are
asked to rate the degree to which they endorse each item on a 4point likert scale. The DASS-21 has been found to be a valid
measure of depression, anxiety, and stress in non-clinical samples (Henry & Crawford, 2004). A web-based version of the DASS21 was employed for this study. Titov et al. (2011) found that the
DASS-21 maintained good internal consistency when used over
the internet, and in this study the pattern of correlations
amongst the scales was very similar to Lovibond and Lovibond
The secondary outcome measure was a program module usefulness rating that participants were asked to complete at the end
of every module. It was composed of single 10-point likert scale
item asking participants to indicate how useful they found the
module. 1 indicated not useful at all and 10 indicated extremely
useful. This measure was simply used to receive feedback from
Sample size
Power analysis was conducted to determine the sample size for
the current study. Using pilot-testing data from this online based
program and an effect size of .80 (Day, Battista, & McGrath, 2010), a
sample of 33 participants in each group was required for a total of
66 participants.


Online-based self-help program

The online-based self-help program is a modied version of a
program originally developed by Currie et al. (2010) for university
students experiencing mild to moderate symptoms of depression,
anxiety and/or stress. The program is based on cognitive-behavior
therapy strategies used to address depression, anxiety and stress.
It consists of 5 core modules (Introduction and Assessment, Activity
and Mood, Motivation, Thoughts and Feelings, and Advanced
Thoughts and Feelings; see Table 1 for description of each module).
Each module is organized in a multimedia workbook format that
includes psycho-education, real-life examples, videos, audio les,
pictures and activities. Participants worked through the modules
sequentially. The average length Mean (SD) of each module was
27.75 (4.8) pages.
Participants were assigned to a program coach (i.e. a trained
graduate or undergraduate student), who contacted them via
telephone or email (based on the participants preference) on a
weekly basis. On average, contact with a coach involved one phone
call or one email a week. Phone calls were on average 15e20 min in
duration, while emails were on average 1e2 pages in length.
The program coaches provided support and encouragement to
participants as well as claried information contained within the
program, but did not provide therapeutic advice to participants
beyond reiterating what was contained within program material.
Coaches also monitored the participant progress and reviewed the
activities completed by participants to ensure the correct use of the
Thorough attention was given to condentiality and data protection. Participants received a username and password after

Table 1
Description of program modules.
Core modules



1. Introduction
2. Activity and mood

Introduction to program features.

Description of emotional distress.
Relationship between activity and mood.

3. Motivation

Building motivation for change.

4. Thoughts and feelings

How thoughts affect feelings.

Identifying and challenging thoughts.

5. Advanced thoughts and feelings

Challenging more persistent

thoughts and core beliefs.

Depression, anxiety and stress questionnaire.

Suicidal ideation screener.
Identifying goals for changing activity and/or avoidance
Decisional balancing chart.
Identifying hierarchal steps to achieve behavioural goals.
Identifying and planning for barriers to change.
Thought records.
Labeling common cognitive distortions.
Challenging negative thoughts.
More practice of challenging thoughts.
Identifying automatic thoughts and core beliefs.
Challenging automatic thoughts and core beliefs.
Depression anxiety and stress questionnaire.

Optional modules
6. Social relationships

The effect of distress on social relationships

and communication skills

7. Stress management

Reducing stressors and managing stress

8. Sleep

Strategies for facilitating sleep

9. Irritability & anger

Managing feelings of irritability and anger

10. Medication

Brief overview of medications for depression

and anxiety.
Strategies for coping with mood related
PMS symptoms

11. Premenstrual syndrome

(PMS) and mood

Analyzing how thoughts and feelings affect

behaviour towards others.
Learning to be assertive.
Strategies to reduce demands and improve work-style.
Deep breathing and muscle relaxation.
Pre-sleep thought record and replies
Progressive muscle relaxation.
Thought records.
Identifying triggers.
Assertiveness and physical relaxation.


V. Day et al. / Behaviour Research and Therapy 51 (2013) 344e351

signing up to the online program. They were encouraged to change

their password after the signing into the program for the rst time.
Only program coaches had access to the online participant data
and all coaches signed a condentiality agreement. Participants
were assigned a subject number and collected data was password
protected and stored on secure servers (SLQ Server 8) devoted to
research data of the Centre for Family Health at the IWK Health
Centre. The servers are protected by several security levels in
addition to being behind the rewall of the IWK Health Centre. All
IT staff members on this project are research employees, are fully
versed in condentiality and sign hospital condentiality
Statistical analysis
Data were analyzed using IBM SPSS Statistics for Mac, version
20, 2011 (SPSS Inc., Chicago, Ill). Completers were dened as participants who completed Assessments 1 and 2. Dropouts were
participants who did not complete Assessment 2. Analyses were
performed on all randomized participants (i.e. Intention to treat
(ITT) analyses). Missing data were imputed using the maximum
likelihood EM algorithm.
To evaluate differences between the randomized groups at
baseline, independent sample t-tests were calculated for continuous variables and Pearsons chi-squares for categorical variables.
To evaluate the efcacy of the online self-help program, data
collected at Assessment 1 was compared to Assessment 2 using a 2
group  2 time points  3 outcomes (i.e. depression, anxiety, and
stress DASS scale scores) mixed-model Analysis of Variance
Secondary analyses included, a 2 time points  3 outcomes
(i.e. depression, anxiety, and stress DASS scale scores) repeated
measures ANOVA to examine depression, anxiety and stress scores
6 months post-intervention. This analysis only included participants who completed the 6-month follow-up assessment (N 24).
Finally, means and SDs were calculated for the self-reported
usefulness ratings of the program modules to determine whether
certain modules were considered to be more useful than others.
Demographic characteristics
See Table 2 for a summary of participant demographic characteristics. Mean age of the sample was 23.55 (SD 4.98,
Table 2
Participant characteristics.

Range 18e45) and 59 participants (89.3%) were females. Participants endorsed depression (Mean 19.97, SD 10.69), anxiety
(Mean 13.76, SD 7.97) and stress (Mean 21.82, SD 7.76)
scores in the moderate level according to normative data provided
by Lovibond and Lovibond (1995). Despite randomization, the
delayed-access group endorsed higher baseline stress scores
compared to the immediate-access group (t(64) 2.365,
p .021). No other differences in demographic variables emerged
between the two groups.
Thirteen participants (19.7%) learned about the study through
posters, 22 (33.3%) learned about the study via email, 21 (31.8%)
learned about the study from the Dalhousie Counselling website,
and 10 (15.2%) through other means (e.g. referral from student
services, advertisement in class). Baseline depression scores were
signicantly correlated with Anxiety (r .39, p .001) and Stress
(r .33, p .006) scores. Baseline anxiety scores were also
signicantly correlated with Stress scores (r .52, p < .001).
Completion rate
Twelve participants (18%) did not complete Assessment 2,
resulting in a completion rate of (80.3%). Nine of the noncompleters were in the Immediate Access Group and 4 were from
the Delayed Access Group. After completing Assessment 2, only 14
(42.4%) participants in the Delayed Access Group went on to
complete the online-based program as well as Assessment 3.
Twenty-four participants (36.4%) completed the 6-month followup assessment.
Intervention effect on depression, anxiety and stress symptoms
ITT analyses of the effect of intervention on participants in the
online self-help program compared with Delayed Access participants rendered a signicant main effect of time on depression
[F(1,64) 57.69, p < .001, h2p . 47], anxiety [F(1,64) 53.39,
p < .001,h2p . 46] and stress [F(1,64) 58.40, p < .001, h2p . 48]
scores. This result indicates that all participants endorsed lower
depression, anxiety, and stress scores in Assessment 2 compared to
Assessment 1 (See Table 3). Signicant group  time interactions
were also found for depression [F(1,64) 4.93, p .03, h2p . 07],
anxiety [F(1,64) 5.65 p .02, h2p . 08]and stress [F(1,64) 8.74,
p .004, h2p . 12] scores. Planned post-hoc analyses, adjusted for
multiple comparisons, revealed that Immediate Access group had
signicantly lower depression (p .044), anxiety (p .021) and
stress (p < .001) scores at Assessment 2 (i.e. after the online selfhelp program) compared with the Delayed Access group. A significant main effect to Stress scores was also found [F(1,64) 16.59,
p < .001, h2p . 21], reecting the initial baseline differences between the Immediate Access and Delayed Access groups at
Assessment 1. Reliable Change Index (RCI) scores were calculated
for all participants to investigate clinical change in depression,
anxiety, and stress symptoms. In the intervention group, 22 participants (67%) demonstrated a reliable reduction (i.e. RCI > 1.96) in
depression symptoms, 26 (79%) in anxiety symptoms, and 17 (52%)
in stress symptoms. In the control group, 12 participants (36%)
demonstrated a reliable reduction in depression symptoms, 13
(39%) in anxiety symptoms, and 8 (24%) in stress symptoms.

Immediate access group

N 33

Delayed access group

N 33

Age mean (SD)

Males (Females)
Year of post-secondary
education mean (SD)
Baseline depression score
mean (SD)
Baseline anxiety score
mean (SD)
Baseline stress score
mean (SD)

24.12 (5.8)
5 (28)
3.75 (2.5)

22.97 (4.0)
2 (31)
2.79 (1.7)

20.06 (11.2)

19.88 (10.4)

13.58 (8.4)

13.93 (7.7)

19.64 (7.2)

24.00 (7.8)

6-Month follow-up

Found out about study


N (%)
7 (21.2)
10 (30.3)
9 (27.3)
7 (21.2)

N (%)
6 (18.2)
12 (36.4)
12 (36.4)
3 (9.1)

6-month follow-up data was available from 24 (36.4%) of participants. A 2 time points  3 outcomes (Depression, Anxiety, and
Stress) repeated measures ANOVA was conducted to examine
whether depression, anxiety and stress scores changed over a 6month period. There were no signicant main effects for

V. Day et al. / Behaviour Research and Therapy 51 (2013) 344e351


Table 3
Mean (SD) of immediate access group (N 33) and delayed access group (N 33) at Assessment 1 (i.e. baseline) and Assessment 2.
Assessment 1


p < .05,

Assessment 2

Immediate access group

Delayed access group

Immediate access group

Delayed access group

Mean (SD)

95% CI

Mean (SD)

95% CI

Mean (SD)

95% CI

Mean (SD)

95% CI

20.06 (11.16)
13.58 (8.36)
19.65 (7.18)


19.88 (10.37)
13.94 (7.69)
24.00c (7.79)


10.43a (4.49)
5.70a (4.49)
11.77b (6.05)


14.60 (9.51)
9.93 (7.77)
20.52 (7.63)


p < .01, cp<.001.

depression [F(1,23) .124, p .73, h2p .005], anxiety [F(1,23) 1.0,

p .33, h2p .04], or stress [F(1,23) .299, p .59, h2p .01].
Planned post-hoc analyses, corrected for multiple comparisons,
revealed that depression, anxiety and stress scores remained stable
for participants 6 months post-intervention (p > .05; See Table 4).
Usefulness ratings of modules
Mean (SD) self-report ratings of the usefulness of each module,
with 1 indicating not useful at all and 10 indicating extremely
useful, were as follows: Module 2: 6.35 (1.54), Module 3: 6.66
(1.71), Module 4: 6.70 (2.28), and Module 5: 7.41 (2.21). Usefulness
data were not available for Module 1 since this was an introduction
and assessment module. * Note: There were no signicant correlations between Usefulness ratings and changes in Depression,
Anxiety or Stress Scores.
This study investigated the efcacy of an online self-help program for university students with depression, anxiety and stress.
The main nding of this study was that participants who had access
to the guided self-help program signicantly improved with
respect to levels of self-reported anxiety, depression and stress,
compared to participants in the delayed access condition. At the
onset of the study, participants endorsed anxiety, depression and
stress symptoms in the high end of the moderate range. After
completing the core program modules, participants endorsed
anxiety, depression, and stress symptoms in the low end of the mild
severity range, suggesting a signicant improvement in these
symptoms. Furthermore, participants who responded to the sixmonth follow-up questionnaire indicated that they maintained
the benets of completing the online based programs with symptoms in the low end of the mild severity range to the normal range
of anxiety, depression and stress.
The effect sizes for improvements in comparison to the control
condition were moderate to large, with the effects for anxiety and
depression being moderate and for stress being close to large. These
are similar to the effect sizes found with guided self-help programs
as reviewed by Andersson and Cuijpers (2009), Coull and Morris
(2011) and Spek et al. (2007). Although in the studies reviewed
by Coull and Morris effects often diminished at follow-up, improvements were fully maintained at follow-up in this study. Spek
et al. (2007) found that effects for anxiety were larger than those for
Table 4
Mean (SD) of DASS scores post-intervention and after 6-months for both immediate
Access and Delayed Access groups who completed the 6-month follow-up (N 24).



6-Month assessment

8.17 (5.90)
5.08 (4.68)
11.17 (6.35)

7.56 (9.30)
6.17 (7.24)
11.92 (7.88)

depression, although this nding was complicated by the studies of

anxiety more often including therapist support, which by itself was
associated with larger effect sizes. The similar effects for depression
and anxiety found within this study appear to support Spek et al.s
(2007) hypothesis that differences in effect sizes between studies of
internet programs for anxiety and studies of internet programs for
depression may have been due to other differences between those
studies in terms of support. It is notable that Titov et al. (2011) also
found similar effect sizes for both anxiety and depression with their
transdiagnostic program. However one complication is that within
the current study a larger effect size was found for stress, which is
conceptually more similar to anxiety and more correlated with
anxiety than with depression as measured by the DASS (Lovibond &
Lovibond, 1995), and indeed within this study initial anxiety and
stress scores correlated (r .52) This particular issue probably will
be claried as programs combining treatment for anxiety and
depression become more common, and more within-study comparisons of changes in anxiety and depression are accumulated.
Interestingly, participants waiting to do the program also tended
to show some improvement. However, these improvements were
not as signicant as those found in the participants who received
the online based program. These results suggest that the on-line
self-help program helped participants learn how to independently reduce their distress and maintained their improvement.
One potential explanation for the improvement in depression,
anxiety and stress symptoms in the wait list condition group is that
people whose distress is somewhat variable over time will tend to
seek help when their distress is higher than average for them (i.e. a
type of regression to mean effect.).
Anxiety, depression and stress DASS scores were signicantly
inter-correlated, and overall showed a very similar pattern of results. This is supportive of the concept of providing a guided selfhelp program which addresses all three problems, at least as long
as there are some options within the program to apply basic CBT
methods to anxiety and/or depression and/or stress, and some
guidance is available about this. Perhaps it is better to consider the
program as self-help for distress, which people may experience
with various mixtures of anxiety, depression and stress.
Although program coaches provided guidance to some participants regarding choices of some module sections related to specically to anxiety, depression or stress, the majority of participants
were able to do this based upon their own initial DASS scores and
brief explanations written within the program. Despite a variety of
choices within the modules, all participants had to complete
modules 1 through 5 consecutively. The authors will be piloting
another version of the program which will permit more choice,
concerning whether to do the more behavioral modules (2 and 3)
or the more cognitive modules (4 and 5) rst.
Participants tended to rate all modules as quite useful. Although
there was a slight apparent trend for ratings of usefulness to be
higher for each successive module, since the number of participants
rating the module was somewhat declining with each module then
it is not appropriate to compare these ratings. Additionally each


V. Day et al. / Behaviour Research and Therapy 51 (2013) 344e351

module explicitly built upon preceding modules, so that the utility

ratings may better conceptualized as reecting the perceived usefulness of the program for the participant up to that point.
Only 61% of participants in the immediate access condition
completed all ve core modules. The fth module, entitled
Advanced Thoughts and Feelings, was designed to provide help
with cognitive reinterpretation in situations that were more
complicated (such as when the triggering situation for an increase in anxiety is noticing initial anxiety) or otherwise difcult
(such as identifying vaguer underlying core beliefs), and it is
possible that this module was not necessary for some participants.
However still only 67% completed 4 modules. Wojtowicz et al. (in
press) have researched predictors of persistence with the same
self-help program as was used in this clinical trial although with a
somewhat different sample, and found that older age and higher
perceived behavioral control predicted persistence. Noncompletion of internet-based self-help programs is a common
problem (Melville, Casey, & Kavanagh, 2010). It could be interpreted
as an inherent risk of a self-help approach (i.e. if people are left
mostly on their own to work on changes, many will make their own
choices about what to do, which may be different than what a
professional would recommend). However it is useful to consider
this within the context that only about half of university students
who come for in-person professional therapy complete therapy as
recommended by their therapist (Lucas, 2012).
There are various limitations to be noted about these positive
results. One is that 89% of the participants were female. The program
was advertised and available to males and females equally, and
within the program some examples were tailored to the sex of the
participant. The student population of the three universities from
which participants were recruited is approximately equal in numbers
of males and females. At the Counseling Centre that serves these three
universities, typically about two-thirds of clients coming for anxiety,
depression and/or stress are female; and this type of proportion was
expected for this program by the authors. As noted in the introduction, epidemiologically anxiety and depression are about twice as
prevalent in female students as male students. Any explanation of an
increased gender difference is only speculative, but one possibility is
that the name of the program Feeling Better may have been
differentially appealing to female students. For other (copyright)
reasons the name of the program has been changed to SHIFT (SelfHelp for Improving Feelings and Thoughts), and perhaps this name
will turn out to be more equally appealing to male and females.
Another important limitation is that participants were university students who took the initiative to volunteer to do the program,
in response to some notice, advertisement and at times a recommendation by a university advisor or other employee. All participants would have been aware, through other notices and the
consent form for the research, that direct free personal counseling
was also available to them on campus. Thus the positive results
were obtained with people who preferred self-help. (Although a
minor caveat to that statement is that some participants may have
chosen to do it while temporarily away from the campus. E.g.
during a summer term or while on some internship.) This is a
common limitation within most of the published literature on
internet-based self-help, in the sense that the results are often
based on media-recruited participants (Coull & Morris, 2011).
Such recruitment and limitation may be very appropriate, since
self-help programs necessarily require more independent initiative
than traditional therapy. However, it has important implications for
how this and other self-help programs are best conceptualized
within the overall context of psychological services to university
and college students. This self-help program is not intended as a
replacement for in-person therapy. Rather, it is an outreach
program for distressed students who are unlikely to come for

personal therapy, either because they prefer just self-help, or

perhaps have some practical limitations that make it difcult to
access in-person therapy.
We suggest that guided self-help programs are best conceptualized as part of a continuum of possible services. At the most
economical level are website-based self-help suggestions and
pure self-help programs, which will appeal to people who want
to remain anonymous, but do require the greatest independent
initiative in applying the suggestions and motivating oneself to
persist with personal changes. At the next level are guided self-help
programs, for which anonymity is not possible but which will still
appeal to people who prefer self-help, and will provide some
assistance in applying ideas and motivation to persist with difcult
changes. After the initial cost of program development, such programs can be relatively economical to provide. E.g. the cost of
providing coaching for the 66 participants in this trial, not counting
duties and costs related to research requirements, was about
$10,000 Canadian, which is about a third of the local cost of
providing therapy by a psychologist if modules are equated to
sessions. Hedman et al. (2011) found that internet-based cognitive
behavior therapy was more cost-effective than in-person cognitive
behavior therapy even when the latter was done on a group basis.
In-person therapy may be more expensive, but perhaps also the
only form of help suitable for people with a strong preference for it,
or who have other characteristics that make it unlikely that they
will independently complete and apply self-help programs. Further
research is needed to clarify which levels of help are appropriate for
which individuals. However, from an ethical perspective of
respecting individuals right to choose which form of help they
want, perhaps it is the ethical responsibility of service providers to
provide people with some choices amongst effective forms of help.
This research was funded by a grant from the Nova Scotia Health
Research Foundation. Dr. McGraths research is supported by Canada Research Chair. Thank you to Melissa Kervin and Susan Battista
for their assistance with subject recruitment and data organization.
Adalf, E., Gliksman, L., Demers, A., & Newton-Taylor, B. (2001). The prevalence of
elevated distress among Canadian undergraduates: ndings from the 1998
Canadian Campus Survey. Journal of American College Health, 50, 67e72.
American College Health Association. (2010). National college health assessment
reference group executive summary.
Andersson, G. (2009). Using the Internet to provide cognitive behaviour therapy.
Behaviour Research and Therapy, 47, 175e180.
Andersson, G., & Cuijpers, P. (2009). Internet-based and other computerized psychological treatments for adult depression: a meta-analysis. Cognitive Behaviour
Therapy, 38, 196e205.
Andersson, G., Estling, F., Jakobsson, E., Cuijpers, P., & Carlbring, P. (2011). Can the
patient decide which modules to endorse? An open trial of tailored internet
treatment of anxiety disorders. Cognitive Behaviour Therapy, 40, 57e64.
Beck, J. S. (2011). Cognitive behavioural therapy: Basics and beyond. New York, NY:
The Guilford Press.
Beck, R., Benedicti, B., & Winkler, A. (2003). Depression and anxiety: integrating the
tripartite and cognitive-specicity assessment models. Journal of Psychopathology and Behavioural Assessment, 25(4), 251e258.
Brown, T., & Barlow, D. (1992). Comorbidity among anxiety disorders: implications
for treatment and DSM-IV. Journal of Consulting and Clinical Psychology, 60,
Carlbring, P., Maurin, L., Torngren, C., Linna, E., Eriksson, T., Sparthan, E., et al. (2011).
Individually-tailored, Internet-based treatment for anxiety disorders: a randomized controlled trial. Behaviour Research and Therapy, 49, 18e24.
Caruso, J., & Salaway, G. (2008). The ECAR study of undergraduate students and information technology, 2008. Educause Centre for Applied Research, Univeristy of
Chambless, D. L., & Ollendick, T. H. (2001). Empirically supported psychological
interventions controversies and evidence. Annual Review of Psychology, 52,

V. Day et al. / Behaviour Research and Therapy 51 (2013) 344e351

Coull, G., & Morris, P. G. (2011). The clinical effectiveness of CBT-based guided selfhelp interventions for anxiety and depressive disorders: a systematic review.
Psychological Medicine, 41, 2239e2252.
Currie, S. L., McGrath, P. J., & Day, V. (2010). Development and usability of an online
CBT program for symptoms of moderate depression, anxiety and stress in postsecondary students. Computers in Human Behavior, 26, 1419e1426. http://dx.
Day. V., Battista, S., & McGrath, P. (2010 May). An introduction to the Feeling Better
program. in Paper presented at atlantic association of college and university student services conference, Saint John, NB.
Eisenberg, D., Golberstein, E., & Gallant, S. (2007). Help-seeking and access to mental
health care in a university student population. Medical Care, 45, 594e601.
Harrar, W. R., Affsprung, E. H., & Long, J. C. (2010). Assessing campus counselling
needs. Journal of College Student Psychotherapy, 24, 233e240.
Hedman, E., Andersson, E., Ljotsson, B., Andersson, G., Ruck, C., & Linderfors, N.
(2011). Cost-effectiveness of internet-based cognitive-behaviour therapy vs.
cognitive behavioural group therapy for social anxiety disorder: results from a
randomized controlled trial. Behaviour Research and Therapy, 49, 729e736.
Henry, J. D., & Crawford, J. R. (2004). The sort-form version of the Depression
Anxiety Stress Scales (DASS-21): construct validity and normative data in a
large nonclinical sample. British Journal of Clinical Psychology, 44, 227e239.
Johansson, R., Sjoberg, E., Sjogren, M., Johnsson, E., Carlbring, P., Andersson, T., et al.
(2012). Tailored vs. standardized internet-based cognitive behavior therapyfor
depression and co-morbid symptoms: a randomized controlled trial. PloS One,
7, e36905.
Kessler, R., Demler, O., & Frank, R. (2005). Prevalence and treatment of mental
disorders, 1990 to 2003. New England Journal of Medicine, 352, 2515e2523.
Kittinger, R., Correia, C. J., & Irons, J. G. (2012). Relationship between facebook use
and problematic internet use among college students. Cyperpsychology,
Behavior, and Social Networking, 15, 324e327.
Kitzrow, M. (2003). The mental health needs of todays college students and recommendations. NAPSA Journal, 41, 167e181.
Krumrie, E. J., Newton, F. B., & Kim, E. (2010). A multi-institution look at college
students seeking counseling: nature and severity of concerns. Journal of College
Student Psychotherapy, 24, 261e283.
Lovibond, S., & Lovibond, P. (1995). Manual for the depression, anxiety and stress
scales (2nd ed.). Sydney: Psychology Foundation of Australia. www.2.psy.unsw.
Lucas, M. (2012). Counselling on campus: client persistence and progress. Journal of
College Student Psychotherapy, 26, 227e240.
Martell, C. R., Dimidjian, S., & Herman-Dunn, K. (2010). Behavioral activation for
depression: A clinicians guide. New York, NY: Guilford.
Melville, K. M., Casey, l. M., & Kavanagh, D. J. (2010). Dropout from internet-based
treatment for psychological disorders. British Journal of Clinical Psychology, 49,


Mitchell, N., & Gordon, P. K. (2007). Attittudes towards computerized CBT for
depression amongst a student population. Behavioural and Cognitive Psychotherapy, 35, 1e10.
Price, E. L., McLeod, P. J., Gleich, S. S., & Hand, D. (2006). One-year prevalence rates
of major depressive disorder in rst-year university students. Canadian Journal
of Counselling, 40, 68e81.
Radhu, N., Daskalakis, Z. J., Arpin-Cribie, C. A., Irvine, J., & Ritvo, P. (2012). Evaluating
a web-based cognitive-behavioral therapy for maladaptive perfectionism in
university students. Journal of American College Health, 60, 357e366.
Robinson, E., Titov, N., Andrews, g., McIntyre, K., Schwencke, g., & Solley, K. (2010).
Internet treatment for generalized anxiety disorder: a randomized controlled
trial comparing clinician vs. technician assistance. PloS ONE, 5, e10942.
Sanchez-Ortz, V. C., Munro, C., Stahl, D., House, J., Startup, H., Treasure, J., et al.
(2011). A randomized controlled trial of internet-based cognitive-behaviorfor
bulimia nervousa or related disorders in a student population. Psychological
Medicine, 41, 407e417.
Scogin, F., Bynum, J., & Stephens, G. (1990). Efcacy of self-administered treatment
programs: meta-analytic review. Professional Psychology: Research and Practice,
21(1), 42e47.
Spek, v., Cuijpers, P., Nyklicek, I., Riper, H., Keyzer, J., & Pop, V. (2007). Internet-based
cognitive behaviour therapy for symptoms of depression and anxiety: a metaanalysis. Psychological Medicine, 37, 319e328.
Tillfors, M., Carlbring, P., Furmark, T., Lewenhaupt, S., Spak, M., Eriksson, A., et al.
(2008). Treating university students with social phobia and public-speaking
fears: Internet delivered self-help with or without live group exposure sessions. Depression and Anxiety, 25, 708e717.
Titov, N., Andrews, G., Davies, M., McIntyre, K., Robinson, E., & Solley, K. (2010).
Internet treatment for depression: a randomized controlled trial comparing
clinician vs. technician assistance. PloS ONE, 5(6), e10939.
Titov, N., Andrews, G., Schwencke, G., Solley, K., Johnston, L., & Robinson, E. (2009).
An RCT comparing effect of two types of support on severity of symptoms for
people completing Internet-based cognitive behaviour therapy for social
phobia. Australian and New Zealand Journal of Psychiatry, 43, 920e926.
Titov, N., Dear, B., Schwencke, G., Andrews, G., Johnson, l., Craske, M.g., et al. (2011).
Transdiagnostic internet treatment for anxiety and depression: a randomized
controlled trial. Behaviour Research and Therapy, 49, 441e452.
Vogel, D. L., Wester, S. R., & Larson, L. M. (2007). Avoidance of psychological
counseling: psychological factors that inhibit seeking help. Journal of Counseling
& Development, 85, 410e422.
Wang, P., Lane, M., & Olfsen, M. (2005). Twelve-month use of mental health services
in the United States: results from the National Comorbidity Survey Replication.
Archives of General Psychiatry, 62, 629e640.
Wojtowicz, M., Day, V. & McGrath. Predictors of participant retention in a guided
self-help program for university students. Journal of Medical Internet Research,
in press.

All in-text references underlined in blue are linked to publications on ResearchGate, letting you access and read them immediately.