The therapist experience of internet delivered CBT for problem gambling: Service integration considerations

Abstract

Introduction

Countries with legalized gambling offer a network of government funded face-to-face therapy, but usage of this expertise is on the decline. One way to address this issue is to recruit therapists from existing services whereby they provide guidance for the delivery of internet delivered CBT.

Objective

To explore the experiences and perceptions of therapists supporting guided online cognitive–behavioural therapy.

Methods

Interviewees were a sub-sample of therapists from a randomised trial that investigated the relative efficacy of online guided self-directed versus pure self-directed interventions in Australia.

Results

In-person, semi-structured interviews with seven service providers were completed, and thematic content analysis identified five themes which related to: participant suitability and screening (e.g., motivation, computer literacy and access); program content and modality acceptability (e.g., amount of content, look and feel); participant information and management (e.g., program engagement and progression); email communication (e.g., use of templates, appointments, rapport building), and; ongoing service integration (e.g., infrastructure, confidence in product). Overall experiences and perceptions of therapists were positive, notwithstanding barriers concerning assessment of participant suitability, participant management systems and low participant engagement.

Conclusions

Key themes emphasized the benefits of Internet-based interventions for problem gambling, and suggested several areas for improvement. Results should inform the development of future treatments to enable flexible tailoring of interventions to individuals.

Highlights

• •Guidance for iCBT can be integrated into routine service delivery.
• •Screening for suitable clients is important to guides.
• •Guidance as it is typically delivered may be constrictive for highly experienced clinicians.
• •Ongoing service integration was viewed positively in terms of developing a suite of services.

1. Introduction

Cognitive-behavioural interventions have been consistently to be effective in the treatment of problem gambling in the short-term (Cowlishaw et al., 2012; Gooding and Tarrier, 2009; Thomas et al., 2011). For problem gambling, CBT includes combinations of a range of components including cognitive restructuring, behaviour substitution, financial regulation, goal setting, exposure, imaginal desensitisation, information provision, problem solving, self-monitoring, relapse prevention and social skills training (Rodda et al., 2018a). Unfortunately face-to-face delivered CBT protocols are also associated with high rates of treatment drop-out (Melville et al., 2007) and most gamblers do not access services due to a range of barriers (Gainsbury et al., 2014; Suurvali et al., 2009). Furthermore, gamblers typically do not present at in-person services until their gambling behaviours have led to multiple complications (Dowling et al., 2015a, Dowling et al., 2015b). Many people access online or telephone-based services, including those who are new to treatment, younger, male, and have a greater amount of online gambling involvement than is seen in in-person treatment settings (Rodda and Lubman, 2014; Rodda et al., 2018b). Unfortunately, there are few evidence-based treatments available in these settings. There is, however, increasing evidence that self-directed interventions are effective for problem gambling, especially if supported by a guide.

A recent taxonomy of the characteristics of gambling interventions reported 35% of RCTs in gambling research (from 1980 to 2016) were self-directed (Rodda et al., 2018a). Of these, just one-third were delivered as internet interventions, with the remainder delivered via traditional workbook. Self-directed interventions that included some form of therapist contact (n = 6) were either a single session or short contact program (i.e., less than four sessions). In these studies, therapists delivered only motivational interviewing (although interventions were described as CBT) face-to-face or by telephone (i.e., no email or online delivery) in an individual setting.

1.1. Guidance for internet interventions

Internet delivered CBT (iCBT) provides a convenient and easily accessible way for gamblers to access treatment, however, studies consistently report high rates of treatment attrition (Christensen et al., 2006). Offering participants limited therapist support in the form of guidance (defined by Andersson et al. (2013) as regular brief messages of encouragement and answers to questions) has been shown to be effective in delivering iCBT for a range of mental health conditions, including depression, anxiety and substance use disorders as well as improving retention rates (Andersson et al., 2013; Hedman et al., 2012b). Indeed, a systematic review examining the impact of guidance on the efficacy of iCBT identified that guided interventions were superior to unguided interventions in terms of improving symptom severity, the amount of contented completed, as well as the number of participants who completed treatment (Baumeister et al., 2014).

Much of the evidence on guidance has been drawn from randomised controlled trials (RCTs), rather than routine service settings. Guidance in these trials is typically provided in a manualised form, and can include providing encouragement, feedback, self-disclosure, answering questions, supporting motivation, advising on the practical aspects of the program and increasing engagement (Andersson et al., 2013; Paxling et al., 2013; Sánchez-Ortiz et al., 2011). Most studies conducted to date provide guidance through weekly scheduled contacts (Andersson et al., 2009; Andersson and Cuijpers, 2008). Although variable, guidance has been defined as consisting of no more than 12 sessions (contacts) with a maximum duration of 20 min each (Cuijpers et al., 2010). In many RCTs, guides are psychology students and, rarely practicing clinical or counselling psychologists (Pihlaja et al., 2018), however Baumeister et al. (2014) reported no difference in participant outcomes according to the level of therapist qualification.

Recently, there has been movement away from efficacy studies and towards the integration of iCBT into existing service systems and routine care. This may include offering blended treatments whereby iCBT is offered before, or in conjunction with, face-to-face treatment (Erbe et al., 2017; Titzler et al., 2018). It also includes the provision of guidance for iCBT which is drawn from existing services and routine care (Erbe et al., 2017; Kivi et al., 2015; Titov et al., 2018). The optimal method for service integration is still being examined and current issues include the incorporation of measurement conducted in both the service and iCBT, selection of therapists (e.g., specialists versus enhancing current skill base), attitudes towards iCBT, participant recruitment (e.g., from the general public or from the service), as well as the best timing of program delivery (i.e., before, during or after in-person treatment) (Andersson and Hedman, 2013). There are also challenges in recruiting face-to-face therapists (inclusive of counsellors and psychologists) to support iCBT in that they may perceive face-to face as a stronger mode of delivery (Bengtsson et al., 2015). Reported barriers to service integration also include lack of time, inadequate knowledge of the program, and skills deficit (Mol et al., 2016).

To date, no studies have attempted to utilise gambling face-to-face therapists as guides for iCBT. In Australia, gambling services are offered free of charge and consist of a network of therapists (inclusive of psychologists, counsellors and social workers). Service usage in Australia has been stable or declined over the past decade (Productivity Commission, 2010), despite a referral pipeline that sees individuals directed to face-to-face services via helplines and online supports. The aim of this paper is to explore the perspective of therapists on online guidance and the feasibility of iCBT integration into problem gambling service delivery. Given this was the first self-directed iCBT program for gambling in Australia, we also aimed to explore their perspectives on the acceptability of the program itself.

2. Methods

2.1. Design of the trial

This research was part of an Australian pilot randomised controlled trial investigating the effectiveness of iCBT for problem gambling (Merkouris et al., 2017). GamblingLess is the first iCBT program with email guidance in Australia that aimed to provide guidance from existing services. The main objective was to evaluate the effectiveness of iCBT with and without guidance provided by therapists within the gambling help service sector. This sector is government funded and provides free face-to-face, online, and telephone counselling, information and support to anyone impacted by gambling problems. Individuals with self-identified gambling problems were recruited from the community from August 2015 to May 2016 and were randomly assigned to iCBT with guidance (n = 101) or iCBT with no guidance (n = 105). Email guidance was provided weekly by therapists from the Gamblers Help treatment network in Victoria, Australia. This trial was registered with the Australian New Zealand Clinical Trials Registry (Trial ID: ACTRN12615000864527). The RCT, along with this qualitative study, was approved by Deakin University Human Research Ethics Committee (Ethics ID: 2014-123) and Eastern Health Human Research Ethics Committee (Ethics ID: E07/2015).

2.2. GamblingLess

The GamblingLess program was developed as a comprehensive and intensive cognitive-behavioural program that emulates the intensity and depth of a face-to-face cognitive-behavioural intervention, and from which brief and targeted interventions can be developed (Merkouris et al., 2017). The program consists of four modules on an optimised website. Program modules were made up of 13–15 activities and each module took approximately 1–2 h to complete. It was expected that each module would be completed over a 1–2 week period. The activity content consisted of motivational enhancement (e.g., negative consequences, values alignment, identification of triggers, goal setting); behaviour modification (e.g., limiting access to money, budgeting, problem solving, relaxation skills, pleasurable activities), cognitive restructuring (e.g., gamblers fallacy, illusion of control, positive expectancies, near misses) and relapse prevention (e.g., urge management, high risk situations, seemingly irrelevant decisions). Activities were delivered with a combination of video, audio, questionnaires, interactive animations, and written activities. At the end of each module participants could answer questions relating to their gambling spend, treatment goals and ability to resist gambling urges as a way of tracking their progress throughout the intervention. There was also the facility to save each activity as a PDF and print for later review. The GamblingLess program allowed participants to complete as many activities as they liked, and in any order they chose. All activities were available to participants from their first engagement with the program. As one of the aims of this study was to explore the acceptability and feasibility of this comprehensive program, with a view to developing a more refined brief program in the future, it was not anticipated that participants would complete all activities in each module. Participants were encouraged to complete one module a fortnight.

A secure portal provided access to online guides to participant demographics and completed post-module surveys. Guidance consisted of one email contact once a week across an eight-week period, an appointment-based email scheduled at the start of the intervention. Participants could email their guide as often as they liked throughout the program, however, the guide would only respond at the scheduled appointment time. All emails exchanged between the guides and their participants were via secure project-specific email addresses. Therapists had access to the overall participant progression through the module and completion of progress scores (measuring gambling spend, treatment goals and ability to resist urges). They did not have access to participant responses to activities or any information on the degree of engagement with the program content. Guidance comprised the provision of support, monitoring progress, clarifying program information, answering technical questions, addressing other problems that arose, and reminders to complete modules. Participants who failed to attend guidance appointments (i.e., did not respond to the weekly email from their guides) were not contacted by other means. Regardless of whether a participant responded, the guide was instructed to continue to send emails during the guidance period.

Participants who were offered guidance in this study (n = 101) were mostly younger than 40 years of age (n = 64, 63.6%), male (n = 66, 65.3%), born in Australia (n = 76, 75.3%) and employed full-time (n = 70, 69.3%). Almost all (95%) were classified as problem gamblers (n = 96) and the most frequent problem activity was electronic gaming machines (n = 76, 75.3%) followed by wagering on horses or dogs (n = 39, 38.6%). Just 17 participants (16.8%) had previously sought face-to-face treatment for problem gambling. Participants were asked to rate, using a 5-star rating scale, the degree of helpfulness of each completed activity. The proportion of participants who completed each activity in the GamblingLess program was calculated using these star ratings (this does not therefore include those who read or reviewed the content but did not complete the interactive activity). Overall, the rate of completion was similar between the unguided (31.4%) and guided (34.6%) intervention groups. In total, 453 emails were sent from guides (M = 5.81, SD = 2.77, range from 1 email to 10 emails) and 36 participants responded with at least one email (range is from 1 to 6). The total emails sent from participants was 73 (M = 1.26, SD = 1.55).

2.3. Therapists (employed as guides)

Seven therapists volunteered to participate in semi-structured interviews (from a total of 11 who provided guidance). Four could not take part due to other work commitments or because they were no longer employed in the service system. Therapists were most commonly female (n = 5) and two were psychologists, three counsellors, one social worker and one provisional psychologist. Three were recruited from regional services and four were from metropolitan services. Six were face-to-face service providers with one recruited from telephone and online services for problem gambling.

Guidance was manualised and supported by a 3-hour training workshop prior to the commencement of guidance. A group peer supervision session was offered part way through the recruitment period (February 2016) which was moderated by a member of the research team. In addition, the guides were provided with ongoing assistance and support from members of the research team across the course of the study, as required. Therapists provided guidance to the program between August 2015 and March 2016 and on average they provided guidance to 9 participants each.

The seven therapists were interviewed by a clinically trained research fellow after completion of the trial (August 2016). In a series of semi-structured interviews, the research fellow used an interview guide to cover three main areas. These were: (1) the acceptability of the program (e.g., How easy was the information in the program to understand? How likely are you to recommend the program to a participant or other clinician?); (2) views on being a guide (e.g., What did you like most about being a GamblingLess guide? What did you like the least about being a GamblingLess guide? Did you feel your role as a guide was helpful to your participants?); and (3) the feasibility of integrating the program into existing services (e.g., How easy or difficult was it to act as a guide for the GamblingLess program, within your current role in your service? Could GamblingLess be offered to participants as part of your normal service delivery? What would stop this from happening or get in the way? What would help or facilitate integration?). Telephone interviews lasting an average of 45 min (range 26 to 86 min) were audio-recorded and transcribed verbatim.

2.4. Data analysis

Thematic content analysis (Braun and Clarke, 2006) was selected as a rigorous but flexible method of understanding the interview data. Data were analysed at a semantic level whereby the focus was on what the therapist said rather than latent meanings. Analysis involved importing transcripts into NVivo qualitative software and reading and rereading transcripts, generating a list of initial categories as codes, collating data relevant to each code, and collating initial codes into potential themes (Braun and Clarke, 2006). Themes were reviewed and discussed with author (SM). Data from the guides' interviews were grouped into thematic domains and representative quotes selected. Quotes have been maintained in their original form (albeit having removed filler sounds such ‘um’ and ‘er’), except where to do so would possibly identify guides or the participants they were working with. For ease of interpretability, gamblers accessing GamblingLess are referred to as participants and guides are referred to as therapists.

3. Results

Overall, there were five themes identified which related to: (1) participant suitability and screening (i.e., demographics, computer literacy and access, education, social isolation, complexity of problems, motivation and commitment); (2) program content and modality acceptability (i.e., amount of content, timing of content delivery, look and feel, perception of effectiveness and active components of iCBT); (3) participant information and management (i.e., program engagement and progression, single page portal, use of own email); (4) email communication (i.e., use of templates, appointments, rapport building), and; (5) ongoing service integration (i.e., infrastructure, confidence in product, timing of program delivery, targeted approach, participant engagement).

3.1. Participant suitability and screening

Therapists reported strong views on the impact of participant suitability to the GamblingLess program. These were associated with demographics including age, location (regional or remote), English proficiency, and education level.

Well, probably the people who are reasonably well educated, reasonably well spoken, are able to articulate their ideas and thoughts well verbally. So probably not people who are in a lot of chaos and distress at that particular time because it requires a lot of cognitive organisation to work through an online program like that.

(Female, 45–50, psychologist)

Internet literacy was also considered important, including good access, technical competence, and confidence in using a device. There was a perception by some therapists that participant suitability would depend on the ability to problem solve and think through issues in order to engage with the online program. Therapists also associated suitability with service barriers including access to treatment, travel (e.g., easy to do at home), and a preference to access help online. This was due to issues related to shame and stigma as well as privacy concerns (e.g., not having printed material at home that other family members might find). However, therapists perceived this kind of program to potentially attract those who were socially isolated. This was viewed as both an opportunity (in terms of it being an easy way for gamblers to connect) and a potential problem (in terms of it being an easy way for gamblers to remain disconnected).

I have to think for a moment. Look, I don't know how to put this, but often we see huge issues in terms of shame and isolation so that by definition, doing something online is a way – it cuts both ways. It's a way to maintain the isolation, because you don't have to go and speak to someone, so whether that's slightly counterproductive in itself, but that's not really what you're asking, I don't think.

(Male, 50–54, social worker)

Conversely, the therapists listed a range of different participants for whom the GamblingLess program may not be beneficial, including those with severe or complex problems (versus moderate problems), or with serious vulnerability, mental health issues or legal problems. Their views indicated that those with complex problems may be in more of a state of crisis or distress and it could be difficult for distressed participants to sit and work through an iCBT program without professional oversight. It is important to note, however, that these perceptions were based on clinical expertise rather than actual presentations of complex client needs as no therapist reported issues associated with crisis or distress.

Therapists also associated suitability with motivation and commitment. There was a perception that it was relatively easy to attract potential participants to the program and have them consent to participate but it was much more difficult to obtain commitment and ongoing engagement.

Somebody with a reasonably high level of motivation to want to make the change because even though there are aspects of the programme that encourage motivation, I think the person has got to come at it with that. I'm not necessarily sure how much, going through the program, would increase the level of motivation to change.

(Female, 45–50, psychologist)

Therapists also commented on shifting motivation throughout the program. They stated that, even though some participants were motivated at the outset, they experienced a range of barriers to change that meant the program was often abandoned for period of time. Although therapists viewed this flexible access to the program as a positive feature, they indicated screening could help identify individuals more likely to comply with the program over time. Some therapists indicated they attempted to maintain this motivation through support and encouragement. Rather than focus on facilitating intrinsic motivation, the focus was on encouragement and offering support to continue.

Keep it upbeat, try and keep it encouraging, try and keep it motivating. Remind them to keep taking it easy and to not assume everything will happen straight away.

(Male, 55–60, social worker)

I just think that the support, maybe the motivation as well, coming from that support person, it's really helpful for them.

(Female, 50–55, psychologist)

Therapists stressed the importance of screening for motivation and commitment was important in order to ensure participants were fully aware of the required involvement.

Screening was also viewed as a way to tailor the program according to suitability (e.g., matching participant characteristics to those demonstrated to be associated with successful engagement and outcomes). This meant therapists wanted more information on who the program worked for so that they could make appropriate recommendations. In particular, they talked about attrition from the first email contact whereby no further contact was made by some participants. Therapists perceived this to be attrition from the program (not just lack of email engagement) and therefore called for screening for suitability.

3.2. Program acceptability

Overall, therapists reported the program as acceptable to people with gambling problems. The overall structure was considered acceptable, however, two therapists noted that it appeared overly long and contained a great deal of content that may be overwhelming for some participants. They commented that face-to-face treatment does not deliver all of its intervention in one sitting. Rather gamblers received the treatment over many weeks. For this reason, there was a perception that modules should open up and become available over time (rather than all at once). In terms of the user experience (i.e., look and feel), all therapists were satisfied, stating that it appeared easy to navigate, simple to understand and comprehensive. They appreciated video and audio content as well as the integration of interactive materials (e.g., use of a slide ruler and open text boxes). Nevertheless, five therapists commented that the modules were overly reliant on text-based content, which was perceived to be a possible barrier to some participants.

I've worked directly with participants in group settings I've been surprised overall by how negative they've found slabs of text but then I don't know if that's applicable to the people that would sign up for this sort of program.

(Female, 50–55, psychologist)

A key to therapists supporting the program was their knowledge of the effectiveness of internet interventions and CBT, as well as how internet interventions worked. Most therapists requested information on the effectiveness of this type of program for gamblers. One therapist expressed scepticism that internet interventions could work in the absence of face-to-face interaction:

I'm a little bit sceptical at one level about whether that format would be therapeutic for people because I'm a face to face counsellor and I think a lot happens when you sit with a person that doesn't happen when a person sits in front of a screen and does something like that. I need to see it work before I'd be confident that it would work.

(Female, 45–50, psychologist)

Another therapist further expressed concerns about the lack of human contact more generally. However, they stated this may be a reflection of being comfortable with the medium.

I think the absence of human contact does concern me. I know a lot of younger people may be more comfortable with it than I am. I'm sort of an older – you know, I'm in my fifties.

(Male, 55–60, counsellor)

Some therapists did not currently deliver CBT and reported different philosophical and theoretical underpinnings. One therapist suggested supporting a program for which they were not familiar posed challenges and could potentially leave participants feeling confused about how to approach their issues.

Therapists were asked if there were areas of the program that could be extended. They stated it would be helpful to build in material addressing social support (e.g., online forums or discussion board), depression, loneliness, anxiety, and substance use, as well as addressing the shame and stigma that is so frequently associated with gambling problems. Usability improvements suggested included redevelopment of the content into more brief and targeted interventions. Also, some felt the program might benefit from additional strategies to enhance motivation and program engagement, such as using mobile phone messaging.

3.3. Participant information and management

The program had been set up so that participants completed a self-evaluation at the end of each module. Unfortunately, most participants did not complete the evaluation, and this was the only information provided to therapists. This led to an assumption by many guides that participants with limited self-evaluation had disengaged with the program. Therapists were interested in knowing more about any activity that participants had attempted, whether that be incomplete or completed activities. Information requested included log-ins, page views, written activities and surveys attempted, and modules completed. Therapists wanted this information to inform email communication so as emails could be tailored according to their client's engagement in the program.

From my perspective, there's not enough integration between what the participant has actually been doing within the program and what information you have, other than if they tell you so in their email exchange.

(Female, 55–60, social worker)

This kind of information was also desired for the purpose of assisting guides with establishing a therapeutic alliance with participants. Almost all of the therapists commented that they felt discouraged or disappointed with the low level of engagement of participants. Having information on what participants were actually doing was perceived as a means of addressing this.

It was highly discouraging to be there and get virtually no feedback, you know, I'm sort of churning out these emails, going in, a certain amount of looking through, looking through, and getting nothing.

(Female, 50–54, counsellor)

In terms of how information was accessed, there was a strong preference for a simple and easy means of access. Therapists requested that more than one click was too many. They requested one system that collated all participant and counsellor activity. In the current study, therapists used four systems for interacting with participants: (a) the program itself; (b) the website hosting participant activity; (c) email system within the guide website; and (d) a spreadsheet for tracking email communication. There was a preference to use their own email to reduce the number of steps and programs involved in communication. Therapists also commented on a preference to receive alerts for new participants as well as participant activity.

3.4. Email communication

Therapists were instructed to provide one email a week that was informed by a template. Therapists reported for the most part they stuck to the templates, delivering general advice, encouragement, acknowledgement, answers to questions, letting the person know that help or support was there if needed. Therapists used pre-developed content (templates) to prepare weekly emails to participants. There was general agreement that templates were not helpful. Therapists perceived them to be less personal and stated that participants perceived them to be computer generated. They suggested that emails should be automated and sent from the system (rather than the clinician). Emails that were appointment based (i.e., sent at the same time each week) were also perceived as restrictive. Instead, therapists wanted to respond immediately (or within 24 or 48 h) and as needed. Therapists would then also engage in much shorter emails that were conversational (rather than instructional) and were focused on the development of the relationship.

I'm not really a huge fan of just sending that really generic, almost automated message because, what's the point in me being there. If you want, it could be easier to set up an automated email service.

(Female, 30–35, provisional psychologist)

There was a perception that participants both liked and disliked email communication. Therapists thought that emails may be annoying or invasive to some participants. Many participants did not respond to emails and in the absence of program engagement information, therapists were left feeling uninformed and unappreciated.

The rejection was terrible. I just thought I'd get a bit back. I'm in the field, because I clearly crave the human interaction, so I wasn't getting my own addiction fed very well.

(Male, 55–60, counsellor)

Therapists did, however, receive positive responses from some participants. This was in terms of acknowledgement that they were there in the background, reduced social isolation, and appreciation of their feedback on progress.

I only got positive responses from participants, and even if there wasn't much it was just thank you, nice to know you're there. It was an acknowledgement that they were appreciating some sort of contact.

(Female, 55–60, psychologist)

I don't think you can ever really underestimate the importance of somebody being there. As good as pure self-help programs are, I still think a therapist or a guide supported program just really has the crucial advantage of having that person being there, should the participant need them.

(Female, 45–50, psychologist)

3.5. Ongoing service integration

Overall, therapists were positive in terms of service integration. They reported that their perceptions of how their services should be delivered had begun to shift away from focusing solely on face-to-face treatments and towards a multi-modal approach. This included email and telephone counselling as well as video counselling. One therapist reported that the infrastructure and protocols were already in place to integrate internet interventions into their service, however, they needed more information on how the intervention would work for their participants including evidence that it would enhance participant outcomes.

Yes, it could be integrated. We'd need much more evidence as I say to be confident that it was a good product that worked for people. And maybe this evaluation will hopefully tell us something about that.

(Female, 45–50, psychologist)

There was a view that any internet intervention could integrate into a suite of services but that there needed to be more emphasis placed on when and how it was delivered. There was a view that it could be offered to people in the early stages of help seeking (as well as those with low- and moderate-risk gambling severity) as a means of engaging them in later treatment. Therapists also suggested that it could be offered post-treatment, especially for those who had a substantial distance to travel or preferred to reduce their treatment frequency/intensity.

Therapists expressed some frustration at the lack of participant engagement, leaving the guides' skills under-used, with involvement that felt more ‘administrative’ than therapeutic (e.g., sending out a weekly email, responding to questions about log-ins). Such role conflict between guide versus therapist was challenging for senior practitioners, such as Therapist 6 who, with 13 years' experience in a regional face-to-face service, found it ‘quite limiting as a guide of what support you could actually offer them to encourage them’ and not to be able to ‘step into counsellor mode.’ This perhaps reflects a clash between the new model of online guidance and background of providing more treatment and interactivity in a face-to-face modality.

Therapists also commented on blended treatment whereby internet interventions are an adjunctive offering. There was a view that the program may have better engagement if supported by a face-to-face therapist. This was in terms of the participant and therapist having already met (or talked on the phone). One therapist stated that they thought people might take the program more seriously if a professional was involved face-to-face. Another therapist stated there may be better engagement if there was an existing relationship. There was concern, however, that it would be yet another product that was sent home with participants that was not engaged with. Overall, however, they viewed the program as useful and helpful.

I think the program's quite useful, but I think it actually needs to be in tandem with something else. So, if it was part of a counselling setting, so you're actually giving them a bit of control themselves, that they can go and do this and we can keep talking about it in counselling, so the counsellor knew what was actually happening, then that would actually be quite useful. I don't know whether it's a stand-alone, because they have to be pretty highly motivated to continue doing it, and clearly, I didn't have many of them.

(Female, 45–50, counsellor)

I think it just needs to be a part of what a service can offer a participant and I think if it is offered to a participant that has an existing relationship with a service then it's probably more likely to be taken up rather than if somebody stumbles over a website.

(Female, 45–50, psychologist)

Overall, six out of seven therapists would consider participating as a guide in the program again, and all expressed appreciation for the opportunity to learn new skills and techniques. The training and supervision for the trial project was generally found to be satisfactory.

4. Discussion

The aim of this study was to explore the perspective of therapists from services providing routine care for problem gambling and their provision of guidance within the context of an iCBT program. Overall, we found feedback to fall within five main themes, namely (i) participant suitability and screening, (ii) program content and modality acceptability, (iii) participant information and management, (iv) email communication and (v) ongoing service integration. The overall experiences and attitudes of therapists towards GamblingLess and iCBT were positive, despite practical problems around assessment of participant suitability and low participant engagement. Although there is limited evidence that iCBT gambling programs are better suited to particular subgroups, there is research demonstrating that gambling severity and demographic variables (e.g., employment status) may have an impact on participant outcomes for reducing social anxiety (Hedman et al., 2012a). To address these issues, future studies investigating service integration need to focus on developing and implementing a screen for suitability as well as a simple but comprehensive participant management system. Guides also raised issues associated with participant retention. To address this issue, iCBT for gambling needs to be better tailored so as to optimise the delivery of effective components while at the same to removing content that does not contribute to improved client outcomes. This is consistent with Carlbring et al. (2011) who suggested tailoring could target those with multiple problems (e.g., panic disorder and depression) by offering common modules together with specific tailored content. For example, it is recognised that there are three distinct profiles of people with gambling problems: (i) those who were subject only to behavioural conditioning; (ii) those who have emotional vulnerability (e.g., mood disorders, history of trauma), (iii) and those who are characterised by risk taking and impulsivity (Blaszczynski and Nower, 2002). Tailoring iCBT on the basis of these three profiles could potentially render the content more relevant, appropriate and concise.

To improve clinician satisfaction, enhanced participant management systems and more freedom in the parameters of therapeutic engagement are required. Future iterations should provide therapists with details of any contact participants make with the program (e.g., page views and log-ins, and completion of activities and screens). This will assist therapists to tailor the content of email contact according to the progress of the individual. However, training on how to manage this content would be required. For example, it is possible that a participant may have completed 1–2 h of content in a week, which would be a lot of information to cover. In this study, guides were provided with templates (i.e., manualised guidance) that could be adjusted according to participant needs. In our study, guides used these as provided (i.e., they made minor adjustments to templates) and perhaps as a consequence templates were perceived as disruptive to the therapeutic alliance. Based on feedback from guides, future studies could investigate automated program emails in conjunction with brief and targeted emails from the guide.

Therapists would prefer to integrate iCBT into their existing treatment approaches and utilise iCBT as a concurrent and/or adjunctive therapy. This is consistent with a Delphi study examining therapist preferences when engaging in a blended approach to iCBT and face-to-face care (van der Vaart et al., 2014) and the lack of flexibility also appears to be a barrier to service integration (Titzler et al., 2018). It suggests that gambling therapists would be open to a blended model given an appropriate model and training. A co-design approach may be warranted, whereby researchers work with services to determine the preferred (and feasible) mode of delivery. This approach may address issues such as the amount and type of therapist engagement (e.g., guidance versus therapy, appointment versus immediate response), timing of iCBT delivery (e.g., before face-to-face treatment, during or post-treatment such as relapse prevention), as well as the mode of support (e.g., email, phone, chat, forums). Furthermore, service integration should ensure guides have training in both the effectiveness of the modality (i.e., internet delivery) as well as CBT more broadly (Shafran et al., 2009).

4.1. Limitations

This is the first study to examine the experiences of therapists acting as guides from a gambling help service sector, but it is not without its limitations. First, we recruited 7/11 (63%) therapists who provided guidance to the project but the sample size for this study was less than optimal. Caution should therefore be applied to these findings because it is unlikely that we reached data saturation. In addition, one of our therapists had previous experience in providing guidance to iCBT and contributed around one-third of quotes cited in this study. This therapist had valuable insight specifically into areas that could be improved. Overall, however, the results were fairly consistent across guides suggesting the issues identified warrant further investigation. Second, therapists involved in this study had volunteered to participate and were likely not representative of the broader problem gambling intervention workforce. Furthermore, the Victorian (Australia) Gamblers Help network has been encouraged by the funder to provide internet delivered services and this may have impacted on their attitudes towards involvement. More information on the attitudes of the broader gambling workforce (in Australia and internationally) is needed in terms of the willingness to recommend and support iCBT in routine care. Second, our broader study comparing guidance versus no guidance found the addition of guidance had little impact on participant outcomes (Dowling et al., 2017). These findings may be because of the model of guidance, lack of engagement in email guidance or the way guidance was implemented. Therapists expressed some discomfort at not being able to use their full range of clinical skills and it is possible that this limited the amount of impact that guides had on participant outcomes. These findings are similar to other studies with therapists and providers in routine care whereby guidance is considered restrictive and an underutilisation of clinical skills (Kivi et al., 2015). A strength of having a face-to-face therapist involved is the potential to develop clinically valuable rapport and engagement. As suggested by Carlbring et al. (2011), a better approach may be to offer an in-person (or video chat) therapeutic encounter at the commencement of treatment. This has some precedent as Titov et al. (2018) successfully integrated iCBT for anxiety and depression across five existing health services whereby the therapist conducted the initial assessment face-to-face (in person or online via teleconferencing) and then provided support via email or other online communication. Finally, our therapists had provided guidance to a limited number of participants. Although there was a view that iCBT could be integrated into existing workloads and that emails should be responded to immediately (rather than appointment based), this view may be different should the number of participant interactions significantly increase.

5. Conclusions

Overwhelmingly, guides viewed iCBT as needed as part of a suite of services that could be provided by the gambling treatment sector in Victoria, Australia. They stated that the sector was ready and willing to support alternative modes of delivery but that there were issues that needed to be addressed in the delivery of guidance in future iterations.

Funding sources

This project was funded by the Victorian Responsible Gambling Foundation, Australia. The Foundation had no role in the study design, data collection, preparation of this manuscript or the decision to submit the article for publication.

Ethical approval

All procedures performed in studies involving human participants were in accordance with the ethical standards of the institutional and/or national research committee and with the 1964 Helsinki declaration and its later amendments or comparable ethical standards.

Declaration of competing interest

The authors have no competing interests to declare in relation to this article. Over the past 3 years, SR, SM, DS, TL, DL, DA, PH, MB and ND have received funding from multiple sources, including government departments and the Victorian Responsible Gambling Foundation. SR, SM and ND have also received funding from the National Association for Gambling Studies (NAGS); a not-for-profit organisation with individual members across all stakeholder groups, which derives its funding from member fees and conference proceeds. SM has been the Victorian state representative (unpaid) on the NAGS Executive Committee (which includes representatives from all stakeholder groups). DL has provided consultancy advice to Lundbeck and Indivior; and has received travel support and speaker honoraria from Astra Zeneca, Bristol Myers Suibb, Janssen, Lundbeck and Servier. None of the authors have knowingly received research funding from the gambling industry or any industry-sponsored organisation.