Not “just for fun”: Gambling, substance use, and the transdiagnostic role of emotion regulation

Abstract

Models of substance use disorders and more recently pathological gambling underscore stress-relief in maintenance of addictive behaviors. This study examines emotion regulation difficulties as predictors of gambling severity in a community sample with and without substance use disorder symptomatology, hypothesizing that more emotion regulation difficulties, and particularly the reliance on avoidance, would be associated with greater addiction severity both for substances and gambling. Adults regular gamblers were recruited using social media advertising for a survey on emotion regulation, gambling, and substance use. As expected, substance use and gambling showed high co-occurrence. Emotion regulation difficulties predicted severity of gambling but not alcohol use symptoms, although correlations were significant for both disorders. Participants with gambling only and comorbid gambling and substance use showed the greatest reliance on emotional non-acceptance and non-awareness. Poor emotion regulation and avoidance of emotional awareness may contribute to the maintenance of addictions, especially gambling pathology. Improvement of emotional awareness, expression, and acceptance may provide a pathway for reducing such behaviors.

Keywords:

• Gambling
• betting
• addiction
• substance use
• emotion
• avoidance

Gambling disorder, previously listed as an impulse control disorder, is now the only behavioural addiction in the Diagnostic and Statistical Manual of Mental Disorders (DSM-5; APA, 2013). This change in taxonomy highlights its relation to substance use disorders (SUDs), which bears relevance for conceptualization and treatment. Potentially similar etiological mechanisms may explain the high levels of comorbidity (Lorains et al., 2011): SUDs prevalence in people with problem gambling is estimated to 28%–50% (Lesieur et al., 1986; Welte et al., 2001), especially for alcohol use disorders (Grant et al., 2015), which can reach over 40% in some USA studies (Cunningham-Williams et al., 1998; Kessler et al., 2008), and was estimated to 16% in Australia (Maccallum & Blaszczynski, 2002).

The two disorder categories share multiple criteria in the DSM-5, including tolerance, loss of control and dysfunction. Both involve difficulty stopping, regulating, or managing the behavior, despite its negative consequences (Grant & Kim, 2001). Both are associated more with male gender, young age, and low socioeconomic status (Rash et al., 2016). Temperamental traits, like neuroticism and impulsivity have been noted as correlates of both (Dash et al., 2019), pointing to shared neurobehavioral pathways, while both show comorbidities with mood and anxiety disorders (Kessler et al., 2008). Given these similarities, an important direction for research is to examine the applicability of well-researched etiological pathways of SUDs, to the conceptualization of gambling pathology, in order to inform therapeutic interventions, and explain comorbidity.

A transdiagnostic risk factor (Aldao & Nolen-Hoeksema, 2010), playing a role in multiple mental disorders, including SUDs, is poor emotion regulation ability, and more specifically the difficulty in recruiting adaptive strategies to regulate intense or aversive emotions. Although no emotion regulation strategy is inherently good or bad, some strategies, if used as long-standing ways of coping, ultimately propagate distress (Karekla & Panayiotou, 2011). These predominantly include forms of avoidance, like behavioural disengagement or experiential avoidance, i.e. the effort not to come into contact with unwanted experiences (Chawla & Ostafin, 2007; Panayiotou et al., 2017). Avoidance is a transdiagnostic risk for many psychological disorders, including SUDs (Papachristou et al., 2018).

To the contrary, being psychologically and physically healthy entail the ability to modulate the valence, intensity, time course, and expression of emotions (Gross, 1998; Tiffany, 1990) flexibly, depending on the situation (Panayiotou et al., 2021). This requires the willingness to accept and tolerate emotions in order to engage in actions important for one’s needs and values. Strategies involving awareness of the problem and one’s affective responses (Esbjørn et al., 2012; Kimhy et al., 2020; Neumann et al., 2010; Shourie & Kaur, 2017), emotional acceptance, problem appraisal, planning, and active coping are generally adaptive for regulating emotions (Karekla & Panayiotou, 2011).

Given the strong connection of both SUDs and gambling with emotional disorders and neuroticism (i.e. frequent experience of intense, negative emotions; Dash et al., 2019), one can postulate that dysregulated emotion plays a role in their onset or maintenance. In turn, unregulated aversive internal experiences, motivate the individual to engage in addictive behaviors in order to cope (Strahan et al., 2011). Unsurprisingly, emotion regulation, and the motive, conscious or unconscious, to escape unwanted experiences has been a central feature in SUDs models, and is recently examined in gambling disorder as well (see reviews by Velotti et al., 2021l; Marchica et al., 2019; Neophytou et al., 2023).

With regard to SUDs, and alcohol addiction in particular, stress-alleviation models, including the self-medication hypothesis by Khantzian et al. (Khantzian, 1997) and the negative reinforcement model of drinking (Baker et al., 2004), suggest that drinking is negatively reinforced by its stress reduction properties (Hawn et al., 2020). Addictive substances can numb emotions, reduce emotion awareness, decrease withdrawal symptoms, and distract from problems and stress. SUDs stress reduction models have received support from evidence that negative affect and various emotional disorders (e.g., PTSD), pre-date the onset of drinking problems, while people with emotional symptoms endorse using substances to cope (Waldrop et al., 2007).

A similar conceptualization has recently been applied to gambling disorder, which has been described as an escape (Weatherly & Cookman, 2014), while Neophytou et al. (2021), found that endorsing gambling as an escape was the gambling motive that most distinguished pathological from recreational gamblers. This does not downplay the role of other motives in gambling and other addictions. In fact, most gamblers will report recreation, socializing, or financial gain as motives for their behavior (Dechant, 2014). Both substance use and gambling may begin as positively reinforcing activities, that produce arousal, excitement, socializing, and entertainment (Francis et al., 2015; Jacobs, 1986), but may end up being sustained by motives to relief withdrawal symptoms, or stress (i.e. a negative reinforcement pathway). Also, there may be individual differences in the motives that are primary. The pathway model of gambling, (Blaszczynski & Nower, 2002) proposes heterogeneous subgroups of problem gamblers, one of which includes emotionally vulnerable people who use gambling to escape. Escape motives may become more dominant as the person acquires tolerance, as per opponent process theories (Solomon, 1980) and as the negative effects, including withdrawal symptoms of SUDs, or social/interpersonal problems created by the addiction itself, accrue. In fact, increased gambling behavior in response to distress, is a DSM-5 criterion for gambling disorder. The need to use substances or gambling to cope may be especially pertinent for individuals with limited access to more effective strategies for emotion regulation, as suggested by the strong emerging association between both SUDs and gambling and poor emotion regulation ability (Berking et al., 2011; Velotti et al., 2021).

1. Current study

Contributing to this line of work, the purpose of the current study, approved by the Cyprus National Bioethics Committee (Ref. ΕΕΒΚ/ΕΠ/2019/100), is to examine the hypothesis that emotion regulation difficulties are associated with increased gambling and SUDs symptoms in the community. It was specifically expected to find that increased levels of both SUDs and gambling symptoms are associated with more avoidant emotion regulation. Furthermore, given the comorbidity between SUDs and gambling pathology, and the association of both disorders with emotional disorders, negative affect and emotion regulation difficulties, it was expected that individuals with high levels of comorbid symptoms for both types of addiction will be especially characterized by escape/avoidance strategies compared to those with symptoms of gambling addiction only.

2. Method

2.1. Participants

The sample was comprised of 478 community adults of whom 412 had complete data for addictions measures and were included in analyses (Males = 274, Females = 138), living in all regions under the jurisdiction of the Republic of Cyprus. Participants were eligible if they were aged > 18 (Mean age = 35.93, SD = 10.39) and engaged in gambling regularly, defined as reporting to spend at least 25 Euros per month on gambling, during the last year. Of the sample, 79% had at least a high-school education, 44.2% had university education, and 79% lived in urban or suburban areas.

2.2. Procedure

Participants were recruited through Facebook and Instagram advertising, using Facebook filters (socio-demographically defined criteria) that targeted prospective participants who lived in the Republic of Cyprus, were ≥ 18 years old, Greek speaking and had Facebook interests such as: Cards, Sky Betting and Gaming, Online Casino, Gambling, Sports Betting, Internet Poker, Betting strategy, Stock Market, Betting in poker, Mobile gambling, Online Gambling, Bingo, Slots, Betting Company, Gambling, Texas Football Holding T, PokerStars, Casino, Betting Sports, Card Games, Zynga Poker, Horse Racing, Poker or Bet365. The advertisement contained the link to the online questionnaire, the first page of which described the study and included electronic consent. Participants who consented and completed the questionnaire were remunerated with a 25 Euro gift card.

3. Measures

The questionnaire packages started with demographics (gender, age, place of residence) and gambling behavior questions, and continued with the following.

South Oaks Gambling Screen Tool-Revised (SOGS): The South Oaks Gambling Screen (SOGS) is a widely used 20-item questionnaire assessing gambling severity (Lesieur & Blume, 1993). Despite some recent criticisms regarding its false-positive detection rates, the SOGS remains the dominant instrument for measuring pathological gambling in research (Goodie et al., 2013; Neophytou et al., 2023), and when used as a continuous report of symptoms, correlates highly with DSM-IV-based structured interview results, to assess gambling severity (Goodie et al., 2013). A cut-off sum score of ≥5 is typically taken to indicate probable pathological gambling. The SOGS demonstrates satisfactory reliability, with Cronbach’s α = .69 in the general population and α = .86 in gambling treatment samples (Stinchfield, 2002). In the present study α = .88.

Alcohol Use Disorders Identification Test (AUDIT): The Alcohol Use Disorders Identification Test (AUDIT) is a 10-item screening tool by the World Health Organization (Babor, 1992) to assess alcohol use and alcohol-related problems, with a cut-off score of ≥8 signifying hazardous levels of drinking. In a Greek sample, Cronbach’s α was .80 (Moussas et al., 2009), and in an undergraduate sample in Cyprus it was .76 (Papachristou et al., 2018). In the present sample, α = .71.

Fagerström Test of Nicotine Dependence (FTND): The 6-item FTND (Heatherton et al., 1991) is a widely used screener assessing nicotine use, with a cut-off score of ≥ 3 indicating some dependence. Seidner, Burling and Burling (2003) report Cronbach’s α across studies between .55-.74. In Cyprus samples reliability was good (Karekla et al., 2010). In this sample α = .63, which falls within the expected range.

Drug Use Disorders Identification Test (DUDIT): The 11-item Drug Use Disorders Identification Test (DUDIT; Bergman et al., 2003) is a screener for substance abuse/harmful use and dependence. Based on the manual, problematic use can be identified at a cut-off score of ≥6 for men and ≥2 for women, however different cutoff scores have been used in various contexts with overall good sensitivity and specificity (Hildebrand, 2015). Because high levels of drug use were not expected in the present community sample (given the generally lower prevalence rates relevant to other European countries; https://www.emcdda.europa.eu/data/stats2022/pdu and the fact that no local epidemiological data exist using the DUDIT to provide local support for specific cutoffs), the clinical cutoff of ≥2 was used for both men and women, aiming for high sensitivity for some problematic use for all participants. Cronbach’s α in previous studies was 0.80 for the total score (Berman et al., 2005). In the current sample α = .87.

Difficulties in Emotion Regulation Scale (DERS)-18: The Difficulties in Emotion Regulation Scale-18 (Victor & Klonsky, 2016) is a brief version of the widely used 36-item instrument (Gratz & Roemer, 2004). The brief version includes the following subscales: Nonacceptance of emotional responses, Difficulty engaging in goal-directed behaviour, Impulse control difficulties, Lack of emotional awareness, Limited access to emotion regulation strategies, Lack of emotional clarity. A total score of emotion regulation difficulties can also be calculated. The DERS-18 shows subscale α ranging between .77 and .90. In this study, for the full scale, alpha = .89.

4. Statistical analyses

Descriptive statistics provided frequencies of each type of addiction in our sample, using the relevant clinical cutoff scores. Bivariate correlations evaluated the association between severity of each addictive behavior and emotion regulation difficulties, looking at symptoms of addictions as a continuum. To address the main hypotheses whether emotion regulation difficulties characterize to a greater degree people with multiple addictions’ problems, participants were separated into four groups—those meeting cutoffs for gambling disorder and at least one SUDs (N = 72; Multiple Addictions group), those meeting criterial for gambling disorder only (N = 97; Gambling only group), those meeting criteria for any or multiple SUDs but not gambling (N = 77; SUDs only group), and those not meeting any cutoffs for addictive behaviors (N = 151; Nonclinical group). Analysis of Variance compared these groups on total DERS-18, and, to delineate if worse emotion regulation is also associated with greater gambling pathology, we compared the groups on total SOGS, self-reported hours spent gambling and money spent. The latter are additional indices of gambling severity, not directly accounted for in the SOGS items. Specifically, time spent gambling was assessed with a single self-report question asking about the hours each week spent gambling. Amount of money spent was assessed using a single self-report likert-type item with five options referring to average sums of money in Euro spent each month, 1 = 1 = 25, 2 = 26–100; 3 = 101–1000; 4 = 1001–10,000; 5 = >10,000. Multinomial logistic regression was then calculated, to identify specific emotion regulation strategies that distinguish the four groups, first with the Multiple addictions group and then the Gambling only group as reference.

5. Results

5.1. Descriptive statistics

Fourty-two percent of the sample met SOGS criteria for probable gambling pathology (score ≥5), an expected effect as the sample was selected for regular gambling. Potential pathological gamblers, i.e. participants with a score of 3 or 4 were 17.5% of the sample, while non-problem gamblers, i.e. those scoring below 3 represented 40.3% of the sample. Also, 4.6% met criteria for some problematic drug use (DUDIT), 32.3% for some smoking dependence (FTND) and 5.6% for harmful alcohol consumption (AUDIT). Of the total sample, 30.8% scored above cutoff for one substance use disorder, 4.4% above cutoff score for two SUDs, and 1% above cutoffs for three SUDs. Of those meeting these cutoffs, with 95 % CI, probable pathological gamblers were 1.390 more likely than non-pathological gamblers to be male, CI(0.913–2.114). The odds ratio for alcohol dependent responders to be male was 1.161, CI(0.466–2.891), for problematic drug users, 1.432, CI(0.501–4.061) and for dependent smokers, 2.459, CI(1.543–3.921), showing as expected an increased risk of males for all addictions.

Regarding comorbidity, with 95% CI, those with probable gambling pathology had 3.34 times greater likelihood of alcohol dependence CI(1.34–8.30), 1.42 times greater likelihood of nicotine dependence CI(0.9397–2.1610), and 2.44 times greater likelihood of problematic use of drugs CI(0.93–6.34), than those not meeting pathological gambling criteria.

6. Correlations among addiction types and associations with emotion regulation

Table 1 shows bivariate correlations among continuous scores on addiction measures, and between each addiction score and emotion regulation difficulties (total DERS-18). Verifying the descriptive statistics above, increasing gambling pathology was significantly related to all other forms of addiction except nicotine. SUDs symptoms were also interrelated.

Table 1. Correlations among addiction scores and emotion regulation

		1	2	3	4	5	6
1	FTND	-
2	AUDIT	.300**
3	DUDIT	.156**	.250**
4	DERS-18 total	.066	.168**	.054
5	Money Gambled	.089	.037	.059	.111*
6	Hours Gambled	.085	.135**	.066	.119*	.389**
7	SOGS	.087	.199**	.168**	.392**	.322**	.430**

Note: FTND = Fagerstrom Test of Nicotine Dependence; AUDIT = Alcohol Use Disorders Identification Test; DUDIT = Drug Use Disorders Identification Test; DERS = Difficulties in Emotion Regulation Scale; Money Gambled = highest self-reported sum of money gambled per week; Hours Gambled = highest number of self-reported hours spent gambling per week.

Alcohol addiction symptoms and gambling addiction symptoms (including SOGS total score, hours and money gambled) were significantly related to difficulties in emotion regulation, while DERS-18 correlations with symptoms of the other two addictions did not reach significance. To clarify the association between gambling, and alcohol addiction symptoms (the two addictions significantly related to emotion regulation) with specific emotion regulation strategies, additional specific correlations were examined. Both gambling (r between .152, p < .01 for difficulties in awareness, and .401, p < .001 for nonacceptance) and alcohol addiction (r between .08, p = .10 for difficulties in awareness, and .187, p < .001 for impulse control) were significantly related to all emotion regulation difficulties, except for the alcohol addiction/awareness effect that was not significant.

7. Group differences in emotion regulation and gambling

Table 2 shows descriptive statistics for the four groups created on the basis of their clinical scores on the different addictions.

Table 2. Descriptive statistics by addiction group

	Multiple addictions	Gambling only	SUDs only	Nonclinical
	Mean (SD)	Mean (SD)	Mean (SD)	Mean (SD)
Age	34.46(10.08)	37.68(10.71)	36.22(10.07)	35.63(10.62)
SOGS	9.03(3.30)	8.84(3.17)	1.97(1.47)	1.49(1.40)
FTND	5.03(2.67)	.09(.36)	4.75(2.05)	.15(.15)
AUDIT	4.17(4.58)	1.33(1.59)	2.32(3.34)	.87(1.31)
DUDIT	1.71(4.75)	.02(.14)	.39(1.37)	0.01(0.01)
DERS-18	49.51(12.88)	47.02(10.40)	40.33(8.44)	38.80(9.40)
Gender composition	M = 56 F = 16	M = 63 F = 34	M = 63 F = 14	M = 78 F = 73

Note: DERS-18 = Difficulties in Emotion Regulation Scale; Nonclinical = participants not meeting screening cutoff criteria for either gambling pathology or any of the Substance Use Disorders assessed; SUDS only = participants meeting screening cutoff criteria for at least one of the Substance Use Disorders assessed but not for gambling pathology; Gambling only = participants meeting screening cutoff criteria for gambling pathology but not for any of the Substance Use Disorders assessed.

On the ANOVAs, the four addiction groups did not differ significantly in age, but differed significantly on all dependent variables reflecting gambling pathology and emotion regulation. For DERS-18 total, F(3, 396) = 25.18, p < .001, η² = .16. Groups also differed in total SOGS, F(3, 396) = 306.42, p < .001, η² = .70, self-reported gambling hours, F(3, 396) = 26.60, p < .001, η² = .17, and money gambled F(3, 396) = 12.71, p < .001, η² = .09. In all cases, the Multiple addictions group and the Gambling only group reported significantly higher difficulties (emotion regulation and gambling pathology) than the Nonclinical and the SUDs only groups, but did not differ significantly between them.

8. Prediction of addictive behavior groups by emotion regulation

The multinomial logistic regression model aiming to identify specific emotion regulation difficulties that distinguish the four groups fit the data significantly, χ²(21) = 131.79, p < .001, showing that emotion regulation strategies explained a significant proportion of variance, between 28.2 and 30.3% based on the Cox and Snell and Nagelkerke statistics respectively. Based on the predictor variables, 44.4% of cases were classified correctly. As seen in the odds ratios (Exp.(B), Table 3), specific strategies significantly distinguished the Multiple addictions group, used as the reference category, from the other groups. Specifically, members of the Nonclinical group were distinguished from the Multiple addictions group by being significantly less likely to report difficulties in emotional awareness, emotional acceptance, and impulse control, while members of the Multiple addictions group were almost 4 times more likely to be male. The Multiple addictions group compared to the SUDs only group was significantly more likely to report difficulties with emotional awareness and impulse control but did not differ significantly in other strategies or gender, while the Multiple addictions group did not differ significantly from the Gambling only group in any emotion regulation strategy or gender.

Table 3. Multinomial logistic regression showing odds of being classified in a specific addictions group, with DERS-18 strategies and gender as predictors

Multiple addictions group as reference
	B		SE	Wald	df	p	Exp(B)	95% CI
								Lower	Upper
Nonclinical	Awareness	−.16	.07	5.02	1	.025	.86	.75	.98
	Clarity	−.10	.09	1.24	1	.265	.91	.77	1.08
	Goals	.01	.08	.01	1	.924	1.01	.86	1.18
	Impulse	−.22	.09	6.14	1	.013	.81	.68	.96
	Nonacceptance	−.20	.09	5.42	1	.020	.82	.69	.97
	Strategies	−.10	.10	.99	1	.321	.91	.74	1.10
	Gender	1.32	.37	12.90	1	.001	3.72	1.82	7.60
Gambling only	Awareness	−.08	.07	1.26	1	.262	.92	.80	1.06
	Clarity	−.02	.08	.09	1	.759	.98	.83	1.14
	Goals	.02	.08	.08	1	.784	1.02	.87	1.20
	Impulse	−.14	.08	3.01	1	.083	.87	.74	1.02
	Nonacceptance	−.105	.08	1.42	1	.234	.91	.78	1.06
	Strategies	.10	.10	1.15	1	.284	1.11	.92	1.34
	Gender	.60	.37	2.69	1	.101	1.82	.89	3.73
SUDS only	Awareness	−.16	.08	4.73	1	.030	.85	.73	.94
	Clarity	.04	.09	.22	1	.642	1.04	.87	1.25
	Goals	.01	.09	.01	1	.945	1.01	.85	1.19
	Impulse	−.22	.10	5.10	1	.024	.81	.67	.97
	Nonacceptance	−.14	.09	2.42	1	.120	.87	.72	1.04
	Strategies	−.12	.11	1.10	1	.295	.89	.72	1.11
	Gender	−.18	.43	.16	1	.687	.84	.36	1.97
	Gambling only as reference
Nonclinical	Awareness	−.08	.06	1.85	1	.174	.93	.83	1.03
	Clarity	−.07	.08	.87	1	.351	.93	.80	1.08
	Goals	−.01	.07	.04	1	.833	.98	.86	1.13
	Impulse	−.07	.08	.83	1	.362	.93	.80	1.09
	Nonacceptance	−.10	.08	1.81	1	.179	.90	.78	1.05
	Strategies	−.20	.09	5.26	1	.022	.82	.69	.97
	Gender	.71	.30	5.83	1	.016	2.04	1.14	3.64
SUDS only	Awareness	−.09	.07	1.71	1	.191	.92	.81	1.04
	Clarity	.07	.09	.63	1	.428	1.07	.91	1.26
	Goals	−.02	.08	.04	1	.836	.98	.85	1.15
	Impulse	−.07	.09	.64	1	.425	.93	.78	1.11
	Nonacceptance	−.05	.08	.29	1	.593	.96	.81	1.13
	Strategies	−.22	.10	4.58	1	.032	.80	.66	.98
	Gender	−.77	.38	4.17	1	.041	.46	.22	.97

Note: DERS-18= Difficulties in Emotion Regulation Scale; Nonclinical = participants not meeting screening cutoff criteria for either gambling pathology or any of the Substance Use Disorders assessed; SUDS only = participants meeting screening cutoff criteria for at least one of the Substance Use Disorders assessed but not for gambling pathology; Gambling only = participants meeting screening cutoff criteria for gambling pathology but not for any of the Substance Use Disorders assessed. Gender coding, 1 = male, 2 = female.

When the Gambling only group was the reference, it was differentiated from the SUDs only and the Nonclinical group by less access to emotion regulation strategies, and from the SUDs only group by being more likely male (see, Table 3).

9. Discussion

This study contributes to identifying risk factors of pathological gambling, by examining the role of emotion regulation, known to be associated with other forms of addiction. As in SUDs models, individuals may engage in progressively heavier gambling, to cope with and escape from unwanted, unregulated emotions. Indeed, according to DSM-5 criteria, gambling becomes more severe during periods of distress. The study also examined whether individuals with comorbid gambling and SUDs would show greater difficulties in emotion regulation than those with each pathology alone, on the assumption that comorbidity may involve more severity and broader emotion regulation dysfunction. We also looked into the specific emotion regulation strategies reported by those with multiple addiction symptoms, SUDs only, gambling only, and those not meeting any clinical criteria. We hypothesized that, similar to many forms of psychopathology (Karekla & Panayiotou, 2011), avoidant strategies, especially non-acceptance of emotions would predict addiction severity.

Study predictions were based on evidence that emotion regulation difficulties constitute a transdiagnostic risk factor across many psychological disorders (Aldao & Nolen-Hoeksema, 2010). The same risk applies to addictive behaviors (Coffey & Hartman, 2008), SUDs (Shadur & Lejuez, 2015) and recently emotion regulation was shown to play a role in gambling (Elmas et al., 2017; Nower et al., 2004; Williams et al., 2012). Furthermore, individuals with addictive disorders show emotional difficulties (Kessler et al., 2008), which may prompt the use addictions for coping.

Hypotheses were mostly supported and the study contributes findings on specific emotion regulation difficulties that can inform interventions for gambling pathology and other addictions. First, descriptive analyses provided verification for the significant comorbidity between gambling and SUDs. As expected, those meeting criteria for pathological gambling were more likely to meet clinical cutoffs for all SUDs examined (nicotine, alcohol, and drug use), consistent with previous evidence of high comorbidity among these addictive behaviors (Lesieur & Rosenthal, 1991; Rash et al., 2016). These disorders were more likely among males than females, as described previously (McHugh et al., 2018), showing that male gender constitutes a risk factor for addictive behaviors. This finding is of relevance to the study hypotheses: It is well-documented that whereas emotional disorders are primarily a female problem (Panayiotou et al., 2021), SUDs are more frequent among males, with a hypothesized explanation that men are socialized not to express emotions and seek help, but may self-medicate with substances (Zalta & Chambless, 2008) for coping and emotion regulation. Gambling potentially serves a similar function.

The high degree of overlap we observed between SUDs and gambling, points to similar etiological and maintenance mechanisms, as hypothesized. An initial indication that emotion regulation may be such a mechanism is seen in the bivariate correlations that show both AUDIT and SOGS were significantly related to emotion regulation difficulties, with the association being strongest for gambling. Such findings are overall consistent with much accruing evidence, as well as important models of SUDs and more recently gambling, which posit a key role of coping motives (Ricketts & Macaskill, 2003). People who become addicted may be less able to cope with negative affect in healthier ways (Kun & Demetrovics, 2010; Marshall-Berenz et al., 2011; Riley & Schutte, 2003; Verdejo-García et al., 2008), either because they are prone to experience intense negative emotions, have emotional disorders, neuroticism or more stressful experiences, and/or have less access to effective, flexible and adaptive emotion regulation and coping repertoires. In fact, lack of access to emotion regulation strategies was found to differentiate the Gambling only group from both SUDs only and non-clinical participants, but not from the Multiple addictions group, in the logistic regression, revealing a specific deficit in coping repertoires among those with gambling pathology. Results are compatible with the “self-medication” hypothesis (Khantzian 1997) and other stress reduction hypotheses that were created to explain SUDs, but appear to hold for gambling as well.

Logistic regression carried out to show specific emotion regulation styles that affect the odds of belong to specific addiction groups are quite informative regarding emotion regulation deficits that should be addressed in relevant treatment and prevention programs. Findings underscored the role of both lack of access to strategies more broadly, and emotionally avoidant approaches in increasing the odd of belonging in the multiple addictions groups in particular. These findings are consistent with research demonstrating how the non-acceptance and intolerance of unwanted emotions perpetuates negative affect and increases risk for psychopathology (Karekla & Panayiotou, 2011). Specifically, the Multiple addictions group was differentiated from Non-clinical participants by non-acceptance and non-awareness of emotions, the two DERS-18 strategies that reflect intolerance of and lack of contact with one’s emotions. Although no causative relation can be posited, findings are consistent with the idea that gambling and SUDs may be reinforced by their function of escape from conscious distress (Gratz & Tull, 2010; Kun & Demetrovics, 2010). Lack of emotional awareness, a prerequisite for initiating effective coping, also differentiated the Multiple addictions from the SUDs only group but not from the Gambling only group, showing that not being in contact with one’s emotional experience is particularly characteristic of those engaging in pathological gambling, irrespective of the presence of other addictions. Lack of emotional awareness has been found to characterize gambling pathology in previous research (Williams et al., 2012). Furthermore, evidence from personality research underscores the role of emotional awareness in addiction: Both SUDs and gambling are characterized by increased alexithymia (Marchetti et al., 2019; Morie & Ridout, 2018), a personality trait reflecting difficulties in identifying and describing emotions, and pervasive emotion dysregulation and avoidance (Panayiotou et al., 2015).

The strong association between gambling and lack of emotional awareness can tentatively be explained by the active nature of this behavior, where one can spend increasingly more time and money on a cognitively demanding activity (relative to the use of substances that often involve automatic action processes; Tiffany, 1990), that may help disengage one from unpleasant experiences. The fact that impulsive action as an emotional regulation strategy also distinguished the Multiple addictions group from both the SUDs only and Non-clinical groups (but not from the Gambling only group) is consistent with this tentative explanation: Gambling was traditionally characterized as an impulse control disorder. Intolerance of distress (non-acceptance) may produce urgency to engage in impulsive action (Schreiber et al., 2012), in order to alleviate undesirable states, preventing the conscious retrieval of efficient, and values-based coping (Berking et al., 2011; Gratz & Roemer, 2004). Even in the context of positive reinforcement motives for addiction, i.e. seeking pleasure, excitement and arousal, positive urgency to impulsively seek reinforcement, despite one’s values and consideration of negative consequences of the behavior, equally represents non-acceptance, of, for example, boredom and lack of excitement that may be reported as justification for addictive behavior (Kim et al., 2019).

Findings of this study, along with the converging recognition that emotion regulation, particularly reliance on avoidance and non-acceptance of emotions increase risk for many forms of psychopathology, can inform decisions regarding therapeutic interventions. Fortunately, emotion regulation represents a malleable target for psychotherapy and psychoeducation. Treatments focused on building emotion regulation skills that include acceptance and emotional expression, like Acceptance and Commitment Therapy, Dialectical Behavior Therapy and more traditional Cognitive Behavioral Therapy (Hayes et al., 2012; Linehan, 2014), targeting recognition and acknowledgement of feelings, expressive writing, emotional communication and learning to seek support, may be particularly effective (Niederhoffer & Pennebaker, 2009; Stanton & Low, 2012). Encouraging psychological flexibility by training the use of a wide repertoire of strategies can increase confidence that something can be done about stressful and emotional circumstances. These strategies can provide a promising alternative to distraction, numbing and avoidance produced by coping with SUDs and gambling.

This study is not without limitations. It represents community data, as it was conducted as part of population studies of the risk factors associated with gambling. As a result, numbers of drug users as measured by DUDIT were small and with low levels of drug use, potentially masking the role of this comorbidity on gambling effects. Gender distribution was uneven between groups meeting clinical criteria and overall males were overrepresented. Also, given the recruitment criterion for gambling, no participants were included with SUDs pathology but no gambling at all, limiting somewhat the generalizability of findings from the SUDs only group. In defense, however, the high comorbidities between addictive behaviors suggest that in the real world, having more than a single addiction is probably the norm rather than the exception. Additional limitations pertain to the selection of screening measures. The SOGS was used for the evaluation of gambling severity, and replication is recommended with the use of other, more clinically oriented psychometric tools that have not been criticized for false detection rates of gambling pathology. It is perhaps the case that the SOGS over-estimated the percentage of pathological gamblers in our sample, which, however is the reason that correlations were provided looking at the measures as dimensional, to highlight associations between addictive behaviors and emotion regulation on the whole sample and not just on those meeting clinical cutoffs. The study, never-the-less would benefit from replication in clinically diagnosed gamblers, as it focused on examining our hypotheses within the community, where no official diagnosis could be made on the basis of screeners, no matter how reliable those may be. Along the same lines, the use of the DUDIT, with a cutoff of 2 for both men and women may be criticized as being low. However, cutoff points vary in different populations and different ones have been used with good sensitivity and specificity (see, Hildebrand, 2015 for a review). It is unclear what the optimal cutoff is for the population in Cyprus, given the absence of epidemiological data, using this tool. Ideally, proper diagnostic tools, including structured clinical interviews must be used in the future to evaluate the overlap of addictive behaviors, within the range of clinical severity. Furthermore, longitudinal research must clarify whether observed emotion regulation difficulties of gamblers represent a stable trait, like alexithymia, or indeed can be effectively trained, in the context of randomized control trials.

Despite any shortcomings, we provide evidence on the role of emotion regulation and specific strategies in addiction, corroborating recent reviews that identify this as a significant risk factor for gambling (Velotti et al., 2021; Neophytou, 2023). The study is in line with efforts to identify transdiagnostic mechanisms that can be addressed through psychoeducational training, to address this common social problem.

Data Availability

Anonymized data will be made available upon reasonable request to the authors. Data cannot be openly shared because permission to do so was not provided by participants.

Disclosure statement

The authors have no interest to declare

Additional information

Funding

This research was funded by a Grant to the first author by the Cyprus National Betting Authority.