The ANU WellBeing study: a protocol for a quasi-factorial randomised controlled trial of the effectiveness of an Internet support group and an automated Internet intervention for depression

This automated online psychological intervention is a research version of the depression stream of the online application e-couch http://​ecouch.​anu.​edu.​au. The intervention comprised a depression literacy module together with five interactive self help tools designed to reduce depressive symptoms.

The depression literacy module contained information about the symptoms and types of depression and their diagnosis, sources of help, information about the prevalence and disability burden associated with depression, depression risk factors, and evidence-based medical, psychological and lifestyle treatments for depression based on updated systematic reviews and clinical practice guidelines. The self help module comprised online versions of treatments amendable to delivery online and known to be effective in face-to-face therapy, bibliotherapy or computerised or Internet delivery. These included: (a) cognitive behaviour therapy; (b) interpersonal therapy; (c) applied relaxation; and (d) physical activity. An example page from the research version of e-couch is illustrated in Figure 1.

The WellBeing Board is a research ISG which uses a bulletin board format to facilitate discussions. It was developed for the trial based on the primary author's experience in running a public ISG and a thematic analysis of a randomly selected subset of topics on that board. The WellBeing board was divided into three main categories: Your Wellbeing, Feeling Better and General. Each week one, or in the case of weeks 8 and 11, two new forums or topics were introduced for each category in turn (see Table 3).

Participants were instructed to login to the website at least twice a week, read any new messages that had been posted, and contribute a minimum of four posts/messages a week. They were also encouraged to login to the WellBeing Board website as often as they wished, to introduce additional topics for discussion and to continue discussions across weeks.

A detailed list of the WellBeing board rules appeared when the participant first registered on the board and was available at all times on the Board home page and in the participant's manual. Discussion of suicide, self harm and other traumatic topics was not permitted. Participants were requested to use pseudonyms and to refrain from posting any material that might identify them. Links and email addresses were filtered out of posts and signatures using an automated script.

The Board was moderated by a member of the WellBeing trial staff who enforced the rules but did not participate in the board discussion. The entry page of the WellBeing Board is illustrated in Figure 2.

The combined intervention involved the delivery of both e-couch and a dedicated WellBeing Board each week over the 12 week intervention period. The trial protocol including reminders was identical to that for the individual conditions. To prevent contamination across conditions separate ISGs were used for the combined intervention and ISG-only conditions. The combined condition ISG also differed from the single ISG intervention condition in that it incorporated an additional forum designed for participants to discuss their experiences on e-couch.

In the absence of a suitable existing web-based attention control condition, the HealthWatch website was purpose built for the WellBeing trial. It comprised 12 modules, each consisting of two online components.

The first component involved a series of online questions probing factors potentially related to depression and wellbeing. This component was modified and extended to 12 modules from a previously reported 5 module telephone control condition [9, 15, 33]. For example, the introduction to the humour topic (Week 11) commenced with the explanation: "Some people believe that humour and laughter can prevent depression and increase the enjoyment of life. Today we would like to ask you to think about humour and whether it is important to your mood". This introduction was followed by a series of 16 multiple choice and open-ended questions: e.g., Do you enjoy telling jokes? (Responses: Yes, No, I can never remember any); How often do you tell jokes (Responses: Never, Rarely, About once a month, at least once a week, Most days); What is the funniest joke you can remember? (Open ended response); Overall do you think your moods are affected by humour?

The second component of the each weekly website module contained online information about a series of 12 topics related to wellbeing. The content in this component of the modules was adapted for Australian use from publications of the National Institute of Health and other US government sources which permit public use. Material was chosen which addressed topics relevant to wellbeing but which contained no or minimal information about interventions for depression or stress.

Participants were instructed to log onto the website each week to respond to the questions in first component of the module, and then to read the material in the second. An example page from HealthWatch website is shown in Figure 3.

As required by ICH guideline E9 [34], randomisation was carried out by the trial biostatistician who was not involved in the day-to-day conduct of the trial.

Potential participants were randomised using a stratified block design with stratification by level of psychological distress (high/low distress level on K10), gender (male, female, uncommitted), and location of residence (metropolitan/rural). Randomization schemes with a fixed block size of four were generated for each strata using the web site Randomization.com http://​www.​randomization.​com.

Depressive symptoms: The primary outcome measure was the Centre for Epidemiological Studies Depression scale (CES-D; [35]). This 20-item self-report scale was used to measure change in the severity of depressive symptoms.

The following secondary measures were used at baseline and in at least one subsequent survey to evaluate improvement or change in mental health, other outcomes and potential moderating or mediating factors:

Anxiety severity: The Penn State Worry Questionnaire (PSWQ)[36]; Quality of life: The EUROHIS QOL - 8 item index [37]; Disability: 'Days Out of Role' questions adapted from the US National Comorbidity Survey [38]; Self esteem: Rosenberg Self-esteem Scale [39]; Perceived social support: The 8-item Medical Outcomes Study Social Support Survey (MOS)/Emotional/informational support subscale [40]. This scale was considered more relevant than others to online as opposed to face-to-face or telephone support. Loneliness: The UCLA Loneliness Scale [41]; Stigma: Depression Stigma Scale (DSS-Personal), a measure of personal, negative attitudes to depression [15, 42]; Depression knowledge: (i) a short 11-item version of the Depression Literacy Scale (D-Lit depression literacy scale; [15]); (ii) 33 depression treatment knowledge items evaluating the perceived helpfulness ('Helpful', 'Neither helpful nor harmful', 'Harmful', or 'Don't know') of three medications, five psychological treatments 21 lifestyle or alternative treatments and 4 formal and informal sources of help [8, 43]; Help seeking: Participants indicated (Yes or No) which of the above medication, psychological, lifestyle and alternative treatments or activities they had used or people they had consulted for depression in the past 3 months [8, 43]; Empowerment: Power-powerlessness subscale of the Empowerment Scale [44]. This scale was developed in conjunction with mental health consumers; Perceived utility of the Internet: Two questions assessed participant's beliefs about the utility of the Internet for preventing and understanding depression ("I am confident that people could learn skills for preventing depression from a website" (Yes, No); "I am confident that a website could help people understand depression better" (Yes, No)).

In addition, the following measures were employed at one or more points post-intervention with a view to assessing their experience of the program: Satisfaction and perceived benefits/negatives: Satisfaction with the interventions was measured using 10 previously employed self-report items (e.g., 'How useful was the website?' (Very useful, useful, Not very useful, Not at all useful)) together with a series of questions based on items developed in a UK study of depression ISG users [45] (e.g., 'The website helped me to discuss subjects that I felt unable to discuss elsewhere.' (Strongly agree, Agree, Neither agree not disagree, Disagree, Strongly disagree) and the Consumer Reports CRES-4 scale [46, 47]; Perceived negative effects were measured with items specifically developed for the purposes of this study (e.g., 'Were any of the following a problem for you as a result of using the website? 'Feeling annoyed or upset by the comments made by other members on the board.' (Yes, No).); Participation in healthcare: A three item measure, modified from a study of a physical health internet support group [48] was employed at post intervention to assess participants' self-reported capacity to discuss treatments with their provider, feel in control and knowledgeable and be assertive with their healthcare provider; Adherence: measured in two ways for each individual: (i) completion of post-test surveys; (ii) number of modules completed (e-couch & Health Watch); (iii) frequency of logins and number of weeks logged onto the WellBeing ISG and activity level on each website (e-couch: number and percentage of exercises completed; number and percentage of pages downloaded; WellBeing Board: number of posts, number of threads started; HealthWatch: number and percentage of questions completed and pages downloaded).

A number of other measures were collected in the screening survey. These included demographic data including gender, age, postcode, marital status, size of town of residence, level of primary and secondary education completed, current study being undertaken, and employment (status, main activity, usual job, past job). In addition, severity of general anxiety, depression and stress was measured using The Depression Anxiety Stress Scale (DASS-21; [49]), social anxiety symptoms were measured using the fear and avoidance subscales of the Brief Social Phobia Scale (BSPS; [50]) and panic symptoms were measured using The 5-item Patient Health Questionnaire (PHQ - Panic; [51]) scale. Participant perceived mental health status (stress, anxiety, panic, social anxiety, or depression), self reported professional help seeking for any such problem and perceived usefulness of seeking professional help for these problems were also collected. A series of purpose designed items were included in the screening survey to assess participant self reported need for information about depression. Finally, an item was included to measure participant randomisation preference at baseline.