Involving concerned significant others in the treatment of problem gambling

Abstract: Background: Problem gambling (PG) is a worldwide phenomenon that affects both gamblers and their concerned significant others (CSOs). Apart from financial difficulties, problem gambling is associated with disrupted relationships, poor mental and physical health, and in some cases criminality. Treatments with cognitive behavioral therapy have generally been efficacious in treating problem gambling, but studies have generally had low adherence to treatment, and few problem gamblers seek treatment. Previous research suggests that involving CSOs can increase treatment seeking, adherence to treatment and improve outcomes for the problem gambler, as well as being beneficial for the CSO. This thesis covers five papers on PG treatment and CSO involvement in PG treatment. Four of the papers build on data from Gambling Free Together (Spelfrihet Tillsammans), an Internet-based randomized controlled study on the involvement of CSOs in PG treatment. The fifth paper covers the results from the Concerned Significant Other Study (Anhörigstudien), an Internet-based randomized controlled trial investigating a support program for CSOs of problem gamblers. Aim: The aim of the present thesis is to study the outcome of Internet-based interventions for problem gamblers and CSOs, the effect of involving CSOs in PG treatment, reasons for drop-out, and the concordance of gambling measures between gamblers and CSOs. Methods: Study I was a parallel-group randomized controlled (RCT) feasibility and pilot trial (N=36, 18 gamblers and 18 CSOs) with 6-month follow up. It primarily investigated the feasibility of the protocol, with the intent of later performing a full scale RCT with the same design. The participants were randomized to either the Cognitive Behavioral Treatment (CBT) condition where only the gambler received treatment, or the condition inspired by Behavioral Couples Therapy (BCT) where both the gambler and the CSO received treatment modules. Participants received 10 treatment modules and weekly telephone support from a therapist. Study II investigated the level of agreement, using intraclass correlation coefficients, between the participating gambler and the respective CSO regarding money lost to gambling, gambling debt and years spent gambling with data from participants in Study I and IV (N=266, 133 gamblers and 133 CSOs). It also included a small simulation study on which assumptions to make regarding distribution of data. Study III was a parallel-group RCT of a support program for CSOs of problem gamblers (N=100), with 12-month follow up. The intervention group received nine treatment modules inspired by Community Reinforcement and Family Training (CRAFT) and weekly telephone support from a therapist, while the other group was put on a waiting-list. Study IV was a parallel-group RCT (that built on the pilot study described above as Study I) investigating the effects of CSO involvement in PG treatment, where participants (N=272, 136 gamblers and 136 CSOs) received 10 treatment modules and weekly telephone support by a therapist. The participants were randomized to either the Cognitive Behavioral Treatment (CBT) condition where only the gambler received treatment modules, or the condition inspired by Behavioral Couples Therapy (BCT) where both the gambler and the CSO received treatment modules. The participants were monitored up to 12 months post-treatment. Study V was a qualitative study on reasons for drop-out from PG interventions. The participants (N=16; 8 gamblers and 8 CSOs) had either dropped out, or were CSOs of gamblers that had dropped out, from Study IV. They were interviewed regarding reasons for drop-out, the process of help-seeking and their views on PG treatment. Results: The results from Study I were inconclusive regarding the relative treatment effect between BCT and CBT gamblers, but both groups significantly improved on all outcomes. BCT was superior than CBT for the CSOs. Feedback from therapists and participants led to some changes to routines and guidelines, but not in treatment content. Results from Study II indicated that CSOs and gamblers had a fair agreement on money lost to gambling, where partner CSOs had a slightly better agreement than other types of CSOs. The simulation study revealed that using a normal distribution when studying money lost might produce unreliable results. Study III showed that CSOs improved on all outcomes related to mental health and relationship satisfaction, but the gamblers’ gambling seemed unaffected. Study IV showed inconclusive results regarding the involvement of CSOs in treatment, but both groups improved on all outcomes. Study V showed that a broad spectrum of themes was related to drop-out, such as changing life circumstances, relapses, and emotional problems, but also a sense of recovery, making treatment seem unnecessary. Conclusions: Supporting CSOs of problem gamblers seem to lower symptoms of psychiatric distress in CSOs, but it is unclear if it has any effect on the gambling of the problem gambler. Involving CSOs in PG treatment might increase the willingness to commence treatment for problem gamblers, but the effects on treatment outcome, and adherence to treatment for the gambler are inconclusive. CSOs seem to have a fair insight into the gambling of the problem gambler, but such insight might vary as a function of the type of relationship. Gamblers dropped out of treatment for various reasons, both because of lack of treatment results but also because they no longer experienced a need for treatment.