Treatment discontinuation in pharmacological clinical trials for gambling disorder

Background Gambling disorder affects 0.5–2% of the population, and of those who receive treatment, dropout tends to be relatively high. Very little is known about participant-specific variables linked to treatment discontinuation/dropout in gambling disorder, especially in pharmacological clinical trial settings. Methods Data were pooled from eight previous randomized, controlled pharmacological clinical trials conducted in people with gambling disorder who had never previously received any treatment for the disorder. Demographic and clinical variables were compared between those who did versus did not subsequently dropout from those treatment trials. Results The sample comprised data from 635 individuals, and the overall rate of treatment dropout was 40%. Subsequent treatment dropout was significantly associated with the following: positive family history of gambling disorder in one or more first degree relatives (relative risk [RR] of dropout in those with positive history vs not = 1.30), preference for mainly strategic vs non-strategic gambling activities (RR = 1.43), lower levels of education (Cohen's D = 0.22), and higher levels of functional disability (Cohen's D = 0.18). These variables did not differ significantly as a function of treatment condition (medication versus placebo). Dropouts and completers did not differ significantly in terms of the other demographic or clinical variables that were considered. Conclusions This study identified several candidate participant-specific predictors of pharmacological treatment dropout in gambling disorder. The findings highlight the need for future studies to address a wider range of contextual variables at large scale (including also study-specific variables e.g. trial/intervention duration), including in naturalistic treatment and clinical trial settings, with a view to developing algorithms that might usefully predict dropout risk.