A Self-Applied Psychological Treatment for Gambling-Related Problems via The Internet: A Pilot, Feasibility Study

Gambling disorder (GD) is a behavioral addiction included in the category of “Substance-Related and Addictive Disorders” of the Diagnostic and Statistical Manual of Mental Disorders (DSM-5). It is characterized by frequent preoccupations with gambling, craving, tolerance, repeated unsuccessful attempts to control or stop gambling, withdrawal symptoms (e.g., irritability or restlessness), gambling to escape from a dysphoric state, “chasing” losses, lying about gambling in significant relationships, and relying on others to finance gambling (APA, 2013). It is a persistent, recurrent pattern of gambling that is associated with substantial impairment (Potenza et al., 2019). For example, the suicide risk in this population is four times higher than in community samples (Wardle et al., 2020), which seems to be explained by certain psychological factors that are characteristic of individuals with GD, such as high impulsivity traits and difficulties regulating emotions (Mallorquí-Bagué et al., 2018). Throughout the paper, we will use the term “problem gambling” when individuals exhibit several symptoms of GD which correspond to 3 or 4 criteria, as well as the term “pathological gambling” when 5 or more criteria are met. These terms are generally used when referring to the Norc Diagnostic Screen for gambling problems (NODS; Becoña, 2004) outcomes.

The complexity of the symptomatology in individuals with GD is often accompanied by other psychological disorders, most frequently anxiety, mood disorders, and substance-use disorders (Cowlishaw et al., 2014; Lorains et al., 2011). It is estimated that the 1-year prevalence of GD oscillates between 0.12% and 5.8% globally (Calado & Griffiths, 2016). In Europe, the yearly prevalence of this disorder ranges from 0.1% to 3.4%. It has been argued, however, that these higher prevalence rates are due to infrequent consultations caused by low awareness of the illness in many patients. These patients often struggle to identify GD symptoms and its negative consequences, finding it difficult to recognize the importance of seeking treatment (Shah et al., 2020).

Encouragingly, there are evidence-based interventions to effectively manage GD. In particular, Cognitive Behavioral Therapy (CBT) is the most frequently used and evidence-based intervention to effectively treat GD, along with motivational interviewing (Menchon et al., 2018; Pfund, et al., 2020; Bodor et al., 2021; Tolchard, 2017). Cognitive restructuring, stimulus control, and exposure with response prevention are the most widely evidenced components of CBT in GD. However, emotional regulation has an essential role in GD and other behavioral addictions (Rogier & Velotti, 2018). Extensions and innovations of CBT have demonstrated their efficacy for substance and behavioural addictions, among which we find Mindfulness-based interventions (Maynard et al., 2018; Sagoe et al., 2021; Sancho et al., 2018); Dialectical and Behavioral Therapy (DBT) adaptations that include mindfulness, emotional regulation, interpersonal effectiveness, and distress tolerance modules that help acquire crucial skills to mitigate emotional dysregulation and maladaptive behaviours through purposeful actions aligned with specific goals and values (Cavicchioli et al., 2020); and interventions that draw upon principles of CBT, including mindfulness, and also on acceptance and positive psychology techniques to help reduce depressive symptoms and gambling-related issues that have shown moderate to strong effect sizes (Bücker et al., 2018). Mindfulness-based therapies enable individuals to not react automatically to emotions of negative or positive valence. For instance, in situations where discomfort arises, such as when experiencing cravings, which can affect one’s inhibitory control capability over gambling behavior (von Hammerstein et al., 2018). Consequently, both CBT and its extensions encompass essential therapeutic components to effectively address GD (Cavicchioli et al., 2020). However, despite the robustness of these interventions when addressing gambling problems, less than 10% of the individuals who suffer a GD seek help, a percentage that is significantly lower than that of other mental health conditions (Gainsbury, et al., 2014; Mojtabai et al., 2002; Suurvali et al., 2008). Also concerning, among those who receive treatment, dropout rates are often high, reaching up to 40% in face-to-face programs (Augner et al., 2022).

As noted earlier, an important barrier to treatment is that individuals with GD are often unwilling to admit that they have a problem and tend to minimize it (Suurvali et al., 2009). They generally seek help when these problems have become extremely severe and have a devastating impact on finances, interpersonal relationships, and physical and mental health (Evans & Delfabbro, 2005; Gainsbury et al., 2014). In addition to this impaired awareness of the problem, there are other barriers that may contribute to the challenges in seeking help. These can be internal, such as fear of stigma, shame, and denial, or external, such as the lack of available or easily accessible services, difficulties in attending treatment sessions due to geographical distance, lack of local expertise and resources, time constraints, and competing work and personal demands (Shah et al., 2020).

As already supported by some research, using Internet-based interventions could help improve access for individuals with GD to evidenced-based interventions. Goslar et al. (2017), for example, reported the effectiveness of two high-intensity structured web-based interventions when compared to face-to-face services for the reduction of problem gambling severity, gambling frequency, and financial loss at post-treatment (Carlbring & Smit, 2008; Casey et al., 2017). In addition, a recent meta-analysis showed that online psychological treatments for GD had moderate effects in the short-term (Augner et al., 2022), with significant positive pooled effect sizes (g = 0.41 for treatment–control comparison, and g = 1.28 for pre-post comparison). Online multi-session treatments also showed larger effects than brief interventions in reducing the amount of time and money spent on gambling (Peter et al., 2019). It is important to note that these self-guided treatments for GD have similar effectiveness when comparing interventions with or without therapist contact, but human contact shows additional advantages in terms of patient satisfaction (Goslar et al., 2017).

In sum, online interventions for individuals with GD appear to be an excellent alternative to make treatments more accessible and scalable. Attrition rates of Internet-based interventions, however, are still an important unresolved issue, with losses that oscillate between 6 and 65% (Bücker et al., 2018; Hodgins et al., 2019; Magnusson et al., 2019). Ecological Momentary assessment and interventions (EMA/EMI), as well as therapeutic support while the online interventions are carried out, are procedures that might help minimize attrition rates in online treatments (Díaz-Sanahuja et al., 2021). The literature on EMIs for individuals with GD is still scarce but encouraging. Hawker et al. (2021), for example, recently conducted a feasibility study with an EMI system to reduce the intensity of craving in people with gambling problems and showed reductions of 71% and 72% in the average number of gambling episodes and craving occurrences, an effectiveness rate that could potentially improve treatment adherence and satisfaction.

Given that there is still very little literature on the use of Internet treatments for individuals with GD enhanced with EMA/EMI, the objective of this study was to assess the feasibility of the “SIN JUGAR, GANAS” [YOU WIN BY NOT BETTING] program, a self-applied psychological online treatment for GD enhanced with EMA/EMI and supported by brief phone-calls. This feasibility, pilot trial will be crucial before conducting a larger-scale randomized controlled trial in terms of potential feasibility problems and preliminary efficacy, which is important for sample size estimation (Aschbrenner et al., 2022). All the previously mentioned aspects will be investigated over the first three treatment modules (i.e., motivation for change, psychoeducation and stimulus control, and responsible debt payment) to evaluate feasibility data before a full-length program is carried out.

This study aimed to show preliminary data about the feasibility (i.e., reach, appropriateness, technology usability, fidelity, and adherence) of the “SIN JUGAR, GANAS” program for inndividuals dealing with gambling-related problems. The study was carried out before conducting a randomized controlled trial to investigate feasibility issues and to describe preliminary effectiveness data of the participants’ progress in managing gambling urges, gambling self-efficacy, anxiety levels, and depressive symptoms. These variables were measured using the web-platform after each module and daily EMA usage. Overall, the feasibility results were encouraging, except for the program’s reach, and preliminary effectiveness findings indicate some improvements, especially regarding the severity of gambling symptoms and self-efficacy to deal with gambling. However, the results were more modest for anxiety and depression.

The sociodemographic and clinical profile of the sample mainly consisted of men (90.9%), with a mean age of 41 years, married or in a relationship, who had completed higher education level studies, and were employed. All the participants were pathological gamblers according to the NODS (Becoña, 2004; Gerstein et al., 1999), which assesses gambling symptomatology during the previous 12 months. Although they were involved in gambling activities since the mean age of 26, they perceived it as a problematic behavior at the mean age of 30 years, after having approximately four years of gambling history. The main forms of gambling behavior were sports betting and slot machines, either online format or in combination with land-based formats. There was a lower percentage of participants gambling only in an onsite format. Most of them (90.9%) reported having financial debts. Our results are in line with previous literature, which supports the representativeness of the sample obtained with our recruitment procedures–which would be positive for reach purposes. The sociodemographic characteristics of our sample are similar to those of Aragay et al. (2021), in which the overall results indicated that participants had a mean age of 45 years, 94.3% were men, they generally had a stable partner, were employed, and indicated an age of gambling onset of approximately 26 years and a gambling history of approximately 5 years. Regarding the type of game and its modality, the most common gambling modes in 2019 in Spain were online sports-betting (31%) and land-based slot-machines (21%) (Dirección General de Ordenación del Juego, 2019; Jiménez‐Murcia et al., 2014), which is again consistent with our sample characteristics.

Online gambling gaining popularity could increase the risk of developing a GD due to its accessibility and the availability of different types of online games (e.g., sports betting, poker, casino games, bingo, and gambling machines) (Aragay et al., 2021; Chóliz, 2016). Sports betting is one of the most prevalent types of game, along with slot machines, with a tendency towards its online format. In our study, most of the participants reported having completed university studies. However, past research examining the profile of participants involved in different types of games found that they typically had only completed primary or secondary education (Aragay et al., 2021). There are probably different types of profiles depending on the type of game participants are involved in. In particular, individuals who bet on sports are more likely to be younger, single, with higher education, and have higher incomes compared to other types of gamblers (Cooper et al., 2022; Dowling et al., 2017; Jiménez-Murcia et al., 2021; Subramaniam et al., 2015). Aragay et al. (2021) also found this specific profile when analyzing only the group of sports betters compared to the group who wagered on land-based slot machines.

In terms of the clinical outcomes profile, our results are similar to previous findings (Zhang et al., 2018), in which approximately 98% of participants suffered from a GD. Several studies indicated that gamblers often have difficulties recognizing their gambling problems (Suurvali et al., 2009) and a high percentage of gamblers sought help when gambling severity was already very high and there was a high impairment or interference in their quality of life (Petry et al., 2005). Aragay et al. (2021) indicated that the overall duration of the GD before the treatment initiation, including online sports betting and land-based slot-machines gamblers, is over five years. Considering only the sports betting gamblers, this period was shorter, which points to different profiles according to the type of gambling behavior. Thus, the sample is representative of the target population according to previous literature (Shaw et al., 2019).

Recruitment and reach difficulties could be influenced by this tendency to ignore the gambling problem until it is very severe. People often seek help when they are in the action stage of readiness to change and experiencing severe symptomatology. This is consistent with our findings, as the sample of the current study consisted of people with GD who were mostly in the action stage. The problem awareness could be associated with the fact that gambling is an acceptable and normalized leisure activity, a means for feeling pleasure and gratification, and GD is defined as a more ego-syntonic disorder (el-Guebaly et al., 2012). For instance, regarding sports-betting activities, there is an established relationship between fun, sports, competition, friendship, and other values associated with youth (Aragay et al., 2021), which probably makes it difficult to recognize when there is problematic gambling (Bijker et al., 2022).

In terms of appropriateness and treatment valuation, the participants had high expectations before starting the intervention. They considered the treatment as logic, they perceived that it would satisfy them, they indicated that they would recommend it to others, they mentioned that it would be useful for the patient’s problem and other problems, and the perceived aversiveness was generally low. These results are in line with other well-established works on the use of online treatments for depression (Romero et al., 2019), which are encouraging the present and future Internet-based treatments for GD.

Regarding fidelity, it was generally not necessary to carry out adaptations on the program’s administration format, the number of modules presented, or the type of support, but we were more flexible with the amount of time allowed for each module and for the duration of the phone-calls. We recommended completing one module per week. The mean duration per module ranged from 4 to 15 h. However, the participants needed a mean of three days to complete the welcome module and the pre-treatment assessment, a mean of five days for module 1 (motivation for change), 15 days for module 2 (psychoeducation), and 13 days for module 3 (stimulus control). Thus, the psychoeducation and stimulus control module took longer than expected. In general, the participants needed about twice as much time as prescribed. However, although they required more time, the welcome and motivation for change module could be completed in less than a week. However, the psychoeducation and stimulation control and responsible debt return modules did need twice as many days to complete. While some previous work did not indicate this need to increase the duration of Internet treatment for GD (Carlbring & Smith, 2008; Myrseth et al., 2013), more research is needed in this area to explore which contexts, for which programs, or for which participants it is important to be flexible with the duration of Internet interventions for GD. In other conditions, for example, emotional disorders, we already have examples of studies recommending the completion of modules over longer periods (i.e., approximately every two weeks) (Mira et al., 2019c), so flexibility might be recommended.

The average adherence to phone-calls was 66.67% and their mean duration was 11.23 min, ranging from 5 to 25 min. This duration was also longer than planned, which affects fidelity. Phone calls took longer when participants presented lapses because the therapist had to make more effort to motivate the participants to continue with the treatment and to avoid gambling again. The duration of phone calls in previous studies regarding GD treatment that included therapist support ranged from 15 to 45 min, so the reported results in the current study are in accordance with previous literature (Carlbring & Smith, 2008; Castren et al., 2013; Myrseth et al., 2013).

Treatment adherence is an important issue to address because treatment dropout rates are high in internet interventions (Pfund et al., 2021). Although the results regarding the contribution of this therapeutic support for self-guided interventions needs more research, some studies report evidence that therapeutic support (e.g., via e-mails, phone-calls, or other channels during therapy) could have a better impact (Petry et al., 2017; Rash & Petry, 2014; Sagoe et al., 2021). Even though the response rates to the phone calls were not always satisfactory, the overall results would support their inclusion in future studies.

In addition to the adherence to the treatment, the adherence to the daily EMA/EMI responses was also modest (i.e., 54.51% of responses provided). These findings are in the same line as Hawker et al. (2021), who found compliance rates for EMA of 51% and EMI of 15%. This suggests that daily evaluation has to be improved, maybe with gamification elements and a sense of utility. Interestingly though, the EMA allowed us to detect that half of the sample did present alarms associated with lapses. In total, 45.5% of the participants reported a mean of two lapses and a range that ranged from 1 to 4, with a mean duration of approximately 3 h and an average money spent of 351.11 € (SD = 388.74; range = 71.53 – 1020.02).The intention of this EMA and the subsequent EMI system was to motivate patients to remain abstinent, or in the event of a lapse, to improve their awareness of the factors that may have affected them, to be able to implement effective strategies in the future in similar risky situations. Future studies could focus on studying the extent to which the intervention affects the motivation to remain abstinent and other gambling-related variables (e.g., gambling urges) in more detail.

Finally, regarding the preliminary results of treatment efficacy, there were significant improvements in the gambling urges and the perceived self-efficacy to cope with gambling urges from pre-treatment to post-module 3. Gambling urges decreased while participants perceived themselves as more capable of dealing with gambling-related high-risk situations. In addition, we also found a non-significant tendency of the participants to improve their anxiety (OASIS; Campbell-Sills et al., 2009; González-Robles et al., 2018) and depressive symptomatology (ODSIS; Bentley et al., 2014; Mira, González-Robles et al., 2019). These are encouraging preliminary results that are in the same line as Hawker et al. (2021), who reported reductions in the average number of gambling episodes, the intensity of gambling urges and frequency, and a rise in gambling self-efficacy over the intervention period using an internet program.

This study has some limitations. First, the sample size was small and we report descriptive preliminary results about the feasibility and preliminary effectiveness of this program concerning only the first three modules. Therefore, these results should be considered with caution. Nevertheless, this study makes substantial contributions by demonstrating the feasibility of continuing this research line, particularly with the involvement of local associations or clinics. The study also showed preliminary evidence regarding the program’s utility. However, it also evidenced the need to adapt the program’s conditions, such as increasing time per module to two weeks approximately and extending the duration of the weekly phone calls when lapses occur. In addition, it is crucial to consider increased therapeutic support, such as proposing a blended format. The online program could be combined with in-person or videoconference group sessions while utilizing EMA/EMI to monitor participants' progress. According to Burlingame et al. (2018), participating in group sessions offers numerous advantages, including opportunities for interpersonal connections, individuality, collaboration, and team dynamics among participants. It also provides individuals with support through shared knowledge and experiences with their peers. Additionally, internet-delivered treatments retain their benefits, such as enhanced accessibility and cost-effectiveness, as they require fewer resources compared to traditional treatments. Although research on blended treatments is still scarce due to their novelty and feasibility, there are still some efficacy results being obtained for the treatment of emotional disorders (Bielinski et al., 2020) and smoking cessation (Choi & Kim, 2018; Siemer et al., 2020). Another limitation refers to the assessment instruments used. In this study, some of them, such as the OASIS and the NODS, showed weak internal consistency, which may be influenced by the sample size and the low variability of the responses. However, previous studies show that they do have adequate psychometric properties (González-Robles et al., 2018; Hodgins, 2004). Likewise, for future research studies, the use of the NODS based on the DSM-5 criteria will be considered, as it has good psychometric properties (Brazeau & Hodgins, 2022). In addition, regarding the limitations of the instruments used, these consisted of self-report measures only. Thus, response bias could be present, which is relevant considering the characteristics of some individuals with GD, who sometimes omit and lie about their gambling behavior to others, as mentioned in the DSM-5 gambling disorder criteria. Thus, although co-therapists are considered in the treatment process to support the patients on the main therapeutic components (e.g., control stimulus and exposure with response prevention), it would be interesting to consider their participation in the EMAs to contrast the self-report information. It would also be interesting to incorporate greater involvement and contact with co-therapists, if possible, allowing them to better understand this issue and how to manage specific situations. This could potentially increase treatment adherence among participants as well. The EMA questions are based on what has happened in the last 24 h. Although it is an assessment of very recent states or behaviors this could still affect the presence of retrospection bias which should be taken into account. Future studies could address these limitations by using event-based or randomly scheduled time-based recordings. However, it’s important to avoid excessive demands on participants to prevent burnout, which was one of the reasons why we only established one measurement time. Another suggestion is the inclusion of more rigorous systems, such as the use of location-based technologies, which would allow evaluation when it detects that individuals are approaching pre-established risk zones.

Finally, the sample was composed of Spanish, Mexican, and Colombian participants who were residing in their countries, except one Colombian participant who lived in the USA. Cultural differences were found such as variations in the language and expressions, in the examples of gambling places, and also in the strategy of self-exclusion included in the stimulus control module. It was necessary to give them similar guidelines within their possibilities given that it was not possible to do self-exclusion as it is done in Spain, as reported by the participants (e.g., resorting to self-exclusion from specific web pages, specific face-to-face venues). Despite these limitations, they were able to follow the program with adequate preliminary results in terms of gambling variables. Ethically, these were people who had a high severity and needed immediate help, and who had difficulties receiving other types of treatment, due to economic limitations and geographical barriers. However, future studies could suggest adjustments for these countries to advance in this research line and better help individuals who suffer from this problem and need help, following, for instance, the guidelines of Salamanca-Sanabria et al., (2019, 2020).

Despite these shortcomings and the recruitment and reach difficulties, likely due to the complexity of GD, the external and internal difficulties in seeking help, and the technological profile of individuals with GD, the current online treatment shows promising preliminary data regarding its excellent appropriateness and usability. Moreover, it offers preliminary data on its generally acceptable adherence and potential utility to reduce gambling symptomatology in people from different Spanish-speaking countries who could not receive help in another way due to geographical barriers, time limitations, and lack of resources, among others. To improve treatment adherence, it may be beneficial to explore modifications such as transitioning to a blended group format and considering the involvement of co-therapists in future research endeavors.