Abstract
Purpose
We conducted a scoping review to map existing research of the potential link between gambling and self-harm. Materials and methods: We searched eight electronic databases and retrieved 190 studies that met our inclusion criteria. Results: Nearly 90% of studies were cross-sectional. Authors most commonly sampled people seeking professional treatment or other help for problem gambling or other substance use/mental health conditions (n = 103, 54.2%). Eighty-two studies (43.2%) provided descriptive analysis only, 50 (26.3%) used a bivariate test to study the direct association between gambling and self-harm, and 58 studies (30.5%) used a multivariate test to account for factors, such as psychiatric conditions, that might partly account for any relationship between gambling and self-harm. Because gambling and self-harm are both multidimensional concepts, we delineated specific categories and mapped the available research according to these categories. We found that most frequently, studies assessed the connection between the presence/severity of problem gambling and suicide attempts (n = 108; 56.8%). Studies of non-suicidal self-harm were rare. When we grouped studies according to both their analytic strategies and their findings, we found that the single largest group of studies (n = 33, 17.4%) were those that reported a positive association between gambling and self-harm within multivariate models. Conclusions: These findings indicate that meta-analytic work, particularly of the association between problem gambling and suicide attempts, is feasible. On a practical level, these findings indicate that people experiencing problem gambling should be screened for self-harm risk.
Acknowledgments
We thank Alex LaRaja for coding and charting the studies in our sample. We thank Karen Amichia for charting the studies. This research was supported by a research contract between the Division on Addiction and GVC Holdings PLC. GVC is a large international gambling and online gambling operator. GVC had no involvement with the development of our research questions or protocol, the data collection/analysis, or development of this manuscript.
Disclosure statement
The authors report no conflicts of interest.
Open Scholarship
This article has earned the Center for Open Science badges for Open Data, Open Materials and Preregistered. The data and materials are openly accessible at https://osf.io/p8df6/, https://osf.io/v3dmz/ and https://osf.io/sh2xr.
Notes
1 Studies conducted prior to the publication of DSM-5 typically used the “pathological gambling” label, but we are using the more current “gambling disorder” label in the text. In our Characteristics of Included Studies table, we report findings using the studies’ original labels.
2 Many recommend replacing the term “suicide gesture” due to its inconsistent usage and arguably dismissive and derogatory tone (e.g., Heilbron et al. Citation2010). We include it here because of its historical use in clinical practice and in research.
3 We did not need to search separately for phrases that begin with “gambl*” or “suicid*,” such as “Gambling Disorder” or “suicide ideation” because such phrases contain the initial word that satisfies the search.
4 Initially, we identified five systematic reviews that might be considered relevant. As we developed the manuscript during the review process, we decided to remove Cheng et al. (Citation2018) because the concept of internet addiction is too distinct from gambling.
5 We used kappa for all characteristics that were coded dichotomously or could be re-coded dichotomously. We used percent agreement for characteristics with more than two categorical response options (e.g., country of first author, minor study design type). We did not attempt to calculate reliability for narrative responses (i.e., sample description, major findings).
6 We coded studies as having any apparent gambling industry funding if any of the listed funding sources were directly part of the gambling industry (e.g., Svenska Spel) or were non-profits funded by the gambling industry (e.g., National Center for Responsible Gaming). Supplemental File 2 shows how we coded each individual funding source.
7 We explain this more fully in the Transparent Changes Document we posted to the Open Science Framework: https://tinyurl.com/u9pu2o6.
8 We included these studies in the “no statistical test” category because although they included a statistical test, it did not examine the direct relationship between gambling and self-harm.