Prevalence of Problem Gambling: A Meta-analysis of Recent Empirical Research (2016–2022)

Abstract

Gambling is widely considered a socially acceptable form of recreation. However, for a small minority of individuals, it can become both addictive and problematic with severe adverse consequences. The aim of this systematic review and meta-analysis is to provide an overview of prevalence studies published between 2016 and the first quarter of 2022 and an updated estimate of problem gambling in the general adult population. A systematic review and a meta-analysis were carried out using academic databases, Internet, and governmental websites. Following this search and utilizing exclusion criteria, 23 studies on adult gambling prevalence were identified, distinguishing between moderate risk/at risk gambling and problem/pathological gambling. This study found a prevalence of moderate risk/at risk gambling to be 2.43% and of problem/pathological gambling to be 1.29% in the adult population. As difficult as it may be to compare studies due to different methodological procedures, cutoffs, and time frames, the present meta-analysis highlights the variations of prevalence across different countries, giving due consideration to the differences between levels of risk and severity. This work intends to provide a starting point for policymakers and academics to fill the gaps on gambling research—more specifically in some countries where the lack of research in this field is evident—and to study the effectiveness of policies implemented to mitigate gambling harm.

Appendices

Appendix 1

Funnel Plots

To assess publication bias, funnel plots were created for the prevalence estimates. As Hunter et al. (2014) recommended for meta-analyses of proportions, sample size (study size) was employed as measure of accuracy on the y-axis.

Funnel plots suggest a certain asymmetry, mainly due to the consistent heterogeneity characterizing the studies, both for cultural, socio-economic issues related to each national/territorial reality, and for reasons related to the different types of sampling, different methods of administration and screening instruments used (Figs. 6, 7 and 8).

Fig.6

Funnel plot—estimates of problem/pathological gambling

Fig. 7

Funnel plot—estimates of problem/pathological gambling without the Mori and Goto study (Japan, 2020)

Fig. 8

Funnel plot—estimates of moderate/at risk gambling

Appendix 2

Meta-regression and Subgroup Analyses

A series of subgroup analyses using a random effects model were performed to investigate possible factors explaining the variability in meta-analytic estimates. First, a meta-regression was conducted to test the overall effect of the following moderators on the mean effect size:

• I. origin (European vs. non-European)

• II. screening instrument (PGSI/CPGI vs other instruments

• III. interview method (telephone interview—CATI or other types; online survey; face-to-face survey—CAPI or other types).

The results of this technique are reported in the table below.

mixed-effect meta-regression—problem/pathological gambling
	Estimate
intrcpt	− 3.8957
metadefPG$OriginNon-European	− 0.3362
metadefPG$MeasurePGSI	− 1.2402
metadefPG$MethodOnline survey	0.7018
metadefPG$MethodTelephone interview (CATI or other types)	0.4512
	Standard error
intrcpt	0.5276
metadefPG$OriginNon-European	0.5017
metadefPG$MeasurePGSI	0.5400
metadefPG$MethodOnline survey	0.4996
metadefPG$MethodTelephone interview (CATI or other types)	0.5028
	p value
intrcpt	< 0.0001
metadefPG$OriginNon-European	0.5028
metadefPG$MeasurePGSI	0.0216
metadefPG$MethodOnline survey	0.1601
metadefPG$MethodTelephone interview (CATI or other types)	0.3696
mixed-effect meta-regression—at-risk/moderate risk gambling
	Estimate
intrcpt	− 3.8019
metadefRISK$OriginNon-European	− 0.5275
metadefRISK$MeasurePGSI	− 0.7642
metadefRISK$MethodOnline survey	0.9893
metadefRISK$MethodTelephone interview (CATI or other types)	1.1865
	Standard error
intrcpt	0.7892
metadefRISK$OriginNon-European	0.5609
metadefRISK$MeasurePGSI	0.7267
metadefRISK$MethodOnline survey	0.5317
metadefRISK$MethodTelephone interview (CATI or other types)	0.5292
	p value
intrcpt	< 0.0001
metadefRISK$OriginNon-European	0.3470
metadefRISK$MeasurePGSI	0.2930
metadefRISK$MethodOnline survey	0.0628
metadefRISK$MethodTelephone interview (CATI or other types)	0.0250

1. Bold parts are the statistically significant data according to the meta-regression

Regarding problem/pathological gambling, the meta-regression did not yield a significant association for the Origin (p = 0.5028) and for the Method (p = 0.1601; 0.3696). Concerning moderate risk/at-risk gambling, meta-regression showed no significant association for Origin (p = 0.3470) and Measure (p = 0.2930).

Below are the analyses for the categories that have yielded a significant association according to the meta-regression: the subgroup analysis by screening instrument for problem/pathological gambling (Fig. 9) and the subgroup analysis by interview method for moderate risk/at-risk gambling (Fig. 10).

Fig. 9

Subgroup analysis for screening instrument (problem/pathological gambling)

The subgroup analysis by screening instrument provides strong evidence of variation. The pooled estimate derived from the PGSI (0.92%; 95% CI 0.57; 1.27) turns out to be significantly lower compared to that of studies employing other screening tools (DSM-IV, SOGS, NODS) (3.04%; 95% CI 0.00; 6.32). This difference can be mainly due to the very high estimate obtained in the study conducted in Japan. There were high levels of between-study heterogeneity in each of these subgroups (I² Measure = PGSI: 96.1% and I² Measure = other instruments: 99.3%).

Fig. 10

Subgroup analysis by methods (moderate risk/at risk gambling)

The subgroup analysis by interview method involves three subgroups: telephone interview (CATI or other types), online survey and face-to-face survey (CAPI or other types). Substantial variations are observed depending on the interview method. In particular, the pooled prevalence of the studies involving face-to-face interview is of a considerably lower magnitude (1.53%; 95% CI 0.40–2.66) compared with the other two interview modes. Studies using online surveys have a value almost twice as high (3.20%; 95% CI 1.45–4.95) compared to face-to-face interviews, while studies using telephone interviews have a high estimate in a middle position between the two modes (2.78%, 95% CI 1.90–3.67). High levels of heterogeneity between studies were also found in each of these subgroups (I² Method = Telephone interview 96.6%; I² Method = Online survey: 97.8%; I² Method = Face-to-face survey: 98.5%).

Below are the analyses for subgroups whose effect was not significant according to the meta-regression (Figs. 11,12 and 13).

Fig. 11

Subgroup analysis by origin—problem/pathological gambling

First, when looking at the world region dichotomy (e.g. European/Non-European prevalence study variable), a higher pooled estimate of the non-European countries (1.64%, 95% CI 0.06; 3.23) is noted compared to the lower result of the European studies (1.06%, 95% CI 0.60; 1.52) (Fig. 11).

Fig. 12

Subgroup analysis by method (problem/pathological gambling)

There is also statistically significant variation on the interview modes. Specifically, the pooled estimate of the studies that used online survey as the method of collection stands out (2.65, 95% CI 0.00; 6.17) (Fig. 12).

Fig. 13

Subgroup analysis by origin (moderate risk gambling/at risk gambling)

There was no evidence of systematic variation in the prevalence estimates by origin. An analysis for the moderator method was not performed because in the case of moderate risk and at risk gambling estimates, only two studies employing instruments other than PGSI are available.