RESEARCH QUESTIONS
Do cognitive-behavioural treatments (CBTs) reduce gambling compared to control groups (i.e., no treatment)? If so, are the reductions maintained over time? Does treatment mode (i.e., individual therapy versus group therapy) impact treatment efficacy?
PURPOSE
Data concerning the efficacy of the treatment of problem gambling, compared to the treatment of other disorders of similar prevalence, is sparse. The purpose of the present study was to carry out a systematic review and meta-analysis (i.e., a systematic approach that integrates and analyzes results from different studies) to determine whether CBTs are effective in reducing gambling behaviour.
HYPOTHESIS
CBTs would lead to greater reductions in gambling than no treatment.
PROCEDURE
The database Web of Science electronic resource was searched for clinical trials published between 1980 and 2008. Studies were included if they: were published in a refereed journal; included a treatment group which consisted, or included a substantial component, of CBTs and included any kind of gambling behaviour or severity measure as an outcome; had a control group as a comparison and/or if provided pre- and post-treatment outcome measures.
INCLUDED STUDIES
Twenty-five studies met the inclusion criteria and were included in the meta-analysis. Nine of the studies were conducted in the US, 8 in Canada, 3 in Spain, and 5 in Australia. The majority (76%) of the studies had a predominantly male sample. The majority (56%) of the studies used DSM-IV criteria for pathological gambling as the indicator of problem gambling behaviour. Nineteen studies provided information on the primary gambling behaviour reported by participants. The majority of studies delivered treatment in individual, one-to-one sessions between a therapist and client. Total treatment hours ranged from 1-148 (with a median of 17) and total number of sessions ranged from 4-112 (with a median of 15). The majority of treatments had been on an out-patient or community basis, and treatment was delivered mainly by psychologists.The therapists were noted to have used a treatment manual or equivalent in 13 of the studies. The studies varied in the number of therapists used, the timing of outcome measurement (i.e., how long after therapy cessation to outcome measures being administered), and the period of follow-up. Twelve studies compared a treatment group with a control group.
OUTCOME MEASURES
Most studies used a number of outcome measures. Frequency and duration of gambling were the most direct measures, along with partial or complete abstinence from gambling. The Clinical Trial Assessment Measure was used to quantify the methodological quality of each clinical study. Only 4 of the studies examined were found to have adequate methodological rigor (i.e., process which produces an outcome that is complete and accurate).
KEY RESULTS
Overall, CBT was found to be effective in reducing gambling behaviours within the first 3 months of therapy cessation (regardless of the type of gambling behaviour practiced) and at 6, 12 and 24 month follow-up periods. Individual and group therapies were equally effective at outcome; however, this equivalence was not clear at follow-up (e.g., only group therapy was effective at 6 months follow-up). All variants of CBTs (i.e., cognitive therapy, motivational interviewing, and imaginal desensitization) were effective; however, there was tentative evidence that when different types of therapy were compared cognitive therapy had an added advantage.
LIMITATIONS
The sample sizes were small for the 12-month and 24-month follow-up periods. Further, there was considerable heterogeneity of the studies included in the analyses.
CONCLUSIONS
Overall, it appeared that CBT (in various forms) had robust enduring short-term effects of reducing gambling behaviours. However, researchers and practitioners should bear in mind that the sample sizes for both the 12- and 24-month time periods were small in size.
