Approximately 2.3% of the world’s population experience problems with gambling (Williams et al. 2012) and for many, gambling problems can evolve into disordered gambling (Wardle et al. 2011). Disordered gambling is classified as a behavioural addiction (American Psychiatric Association 2013) and is characterised by a condition that leads to family, social, personal or recreational pursuits being compromised, disrupted or damaged by continuing maladaptive behaviour (Delfabbro 2013). The disorder often results in serious negative consequences including illegal activities and crime (Lahn 2005), mental health problems (Lorains et al. 2011), financial crises (Grant et al. 2010), domestic violence, (Roberts et al. 2016, 2018), homelessness (Sharman et al. 2015) and self-harm/suicide (Ledgerwood and Petry 2004), and tends to cluster with other high-risk behaviours such as alcohol and substance misuse (Petry 2005).

Several studies have shown that prisoners tend to be impulsive and poor decision-makers in risky situations with a reduced sense of regret (Hughes et al. 2014; Pachur et al. 2010; Varghese et al. 2014; Yechiam et al. 2008). This may explain not only participation in criminal behaviour but also why the prevalence of gambling is much higher than in the general population; as disordered gambling severity escalates, illegal behaviour has also been shown to increase (Lahn 2005). The prevalence of disordered gambling in forensic populations is high (Lai 2014; Turner et al. 2013; Williams et al. 2005). In a recent UK study, the prevalence of disordered gambling in prisons was found to be significantly higher than in the general population at 12.1% (compared with 0.7% (Wardle et al. 2011)) (May-Chahal et al. 2017). International studies have reported even higher prevalence rates; in New Zealand, Abbott et al. (2005) reported that 21% of prisoners were probable lifetime disordered gamblers; and in Australia, it has been acknowledged that as many as 60% of prisoners have a lifetime gambling disorder with 18% reporting they were incarcerated due to offending relating to their gambling problem (Riley et al. 2018). In a study conducted in Canada, up to 65.2% of severe problem gamblers reported gambling-related criminal behaviour (Turner et al. 2009). The majority of perpetrated crimes within disordered gamblers are motivated by gambling either to finance gambling activities or to pay off gambling debts (Rosenthal and Lesieur 1996; Zorland et al. 2008).

Criminal activity is also common within non-forensic clinical gambler populations. Previous research suggests that almost half of all disordered gamblers participate in criminal activities (Zorland et al. 2008). For example, in North America, offending behaviour was assessed in a sample of problem gamblers who contacted a gambling helpline and 43.7% reported committing criminal acts directly related to gambling (Potenza et al. 2001). Likewise, Lorenz et al. (1990) found that 61% of problem gamblers in Maryland had engaged in gambling-related illegal activities. In studies of 60 Gamblers Anonymous attendees in Canada, 68% of gamblers indicated that they had engaged in illegal activities, and 17% of gamblers had appeared in court as a consequence of gambling-related crime (Ladouceur et al. 1994). In Germany, 54.5% of attendees at a gamblers anonymous group indicated they had engaged in illegal activities, and 46% of attendees disclosed committing crimes entirely to support gambling (Meyer and Fabian 1992). Folino and Abait (2009) found that 32% callers to a gambling helpline admitted participating in some sort of illegal exploit due to gambling and this number rose to 76% in those attending Gamblers Anonymous (GA). In the UK, one study identified the proportion of both reported and unreported crimes amongst gamblers attending GA was between 77 and 82% (Brown 1987). Furthermore, reported prevalence rates may be underreported as many of the crimes remain either undetected (Adolphe et al. 2019) or may be committed against persons unlikely to press charges (e.g. family members) (Sakurai and Smith 2003).

It has been suggested that crime may be an indicator of the severity of disordered gambling (Folino and Abait 2009; Bowden-Jones 2013). Previous studies have found more severe lifetime and recent gambling symptomology in those with offending behaviour (Ledgerwood et al. 2007). Likewise, it has been recognised that gambling-related criminality is usually seen in patients with the most severe clinical profiles. Specifically, problematic drug and alcohol use have been associated with an increased likelihood of both engagement in illegal behaviours and the risk of gambling-related arrest (Potenza et al. 2000, 2005). Similarly, problem gamblers with a history of arrest are more likely to have antisocial personality disorder (Potenza et al. 2000) and depression (Momper et al. 2010), and individuals who engage in gambling-related crime are more likely to have experienced suicidality (Ledgerwood et al. 2005).

Despite the global expansion of commercial gambling environments and documented associations with criminality, no relevant studies of clinical populations comprising individuals attending a problem gambling (outpatient) treatment clinic in the UK have been reported as far as we are aware. Evidence for the association between gambling and criminality mainly comes from forensic populations (e.g. Abbott and McKenna 2005; Abbott et al. 2005; May-Chahal et al. 2017; Preston et al. 2012; Turner et al. 2009), with considerably fewer studies carried in clinical populations, which are outside the UK (e.g. Mestre-Bach et al. 2018a, b; Blaszczynski and McConaghy 1994). Moreover, very few clinical studies have asked whether the criminal activity was instrumentally carried out to finance gambling.

The broad aim of this study was to better understand the relationship between gambling and criminality by comparing those who have offended to finance gambling (in a lifetime) to those who have not, in a large sample of individuals seeking treatment for gambling problems. More specifically, the study aimed to:

  1. 1)

    Examine the prevalence of criminal behaviour committed to finance gambling, among both males and females seeking treatment at a clinic for gambling disorder in the UK; and

  2. 2)

    Establish whether clients reporting such offending behaviour could be distinguished from those that did not in terms of a range of clinical characteristics to signal vulnerabilities to disordered gambling; it was predicted that criminal behaviour to finance gambling would be associated with a more severe clinical profile. More specifically, those with a criminal history would score more highly on the PGSI, the PHQ-9 and GAD-7, indicating an increased severity of gambling, depression and anxiety.

  3. 3)

    Evaluate whether clients reporting such offending behaviour could be distinguished from those that did not in terms of gambling-related variables. It was predicted that criminal behaviour committed to finance gambling would be associated with more extreme gambling-related activity (i.e. higher expenditure and time spent gambling, higher debt, a history of parental gambling and earlier age of onset) and more damage and disruption though gambling to relational and professional pursuits.

Method

Sample

The study was based on routine data collected from individuals seeking treatment at the National Problem Gambling Clinic (NPGC), part of the Central North-West London National Health Service (NHS) Trust, UK, from 2011 until 2015 (4 years). The sample of n = 1226 comprised 1128 males (92.0%) and 98 females (8.6%), with an average age of 35.6 (SD = 10.9). Clients were referred from a number of different sources including self-referral, family members, general practitioners and other healthcare professionals. The only inclusion criterion was that patients were classified as problem gamblers (scoring 8+ on the PGSI–see below). The demographic characteristics of the sample are shown in Table 1.

Table 1 Demographics
Full size table

Measures

Problem/Pathological Gambling

Problem Gambling Severity Index

The PGSI is a 9-item self-report tool that measures gambling problem severity (Ferris and Wynne 2001). The scale assesses problematic gambling behaviour alongside interrelated adverse consequences and harms. The items are scored from 0 to 3 (0 = never, 1 = sometimes, 2 = most of the time, 3 = almost always). The PGSI is scored out of 27. The scale has good internal reliability (Cronbach’s α = 0.84) (Ferris and Wynne 2001), and more recent studies have reported excellent internal consistency reliability (Cronbach’s α = .90) (Williams and Volberg 2014).

Offending Behaviour

Participants were asked whether they had committed any illegal behaviours to finance gambling. This included unreported crime, crime reported to the police and any previous convictions/sentences. Offending behaviour to finance gambling was coded as positive when participants answered yes to any part of the question.

Previous Mental Health Diagnosis

Participants were asked if they had even been diagnosed with a mental health diagnosis. This included substance misuse, psychosis, mood disorders, personality disorder or a learning disability/developmental disorder.

Current Depression

Patient Health Questionnaire-9

The PHQ-9 is a widely used 9-item instrument which scores each of the 9 DMS-IV criteria for depression (Kroenke et al. 2001). Items are scored from 0 to 3 (0 = not at all; 1 = several days; 2 = more than half the days; 3 = nearly every day). Scores range from 0 to 27. It has been commended for its high sensitivity and specificity for assessing severity of depression in several populations with a Cronbach’s α of 0.89 and 0.86 (Kroenke et al. 2001).

Current Anxiety Disorder

Generalized Anxiety Disorder-7 Item Scale

The GAD-7 is a widely used 7-item brief measure of generalized anxiety disorder that has been found to have good sensitivity and specificity for GAD (Cronbach α = .92; Spitzer et al. 2006). Questions are scored from 0 to 3 (0 = not at all 1 = several days; 2 = more than half the days; 3 = nearly every day). Scores range from 0 to 21 (Spitzer et al. 2006).

Monetary Debts and Losses

Participants were asked questions about their estimated current, monthly and total debts through gambling using an open response format.

Childhood Trauma

Clients were asked if they had ever experienced childhood abuse. This included physical, sexual and emotional abuse. Multiple abuses were coded when clients indicated that they had experienced all three types of abuse.

Gambling History

Clients were asked the age they first gambled, whether there was a history of parental gambling, and whether they had an early big win. They were also asked to estimate the number of days gambled in the past month, to confirm whether they had previously received treatment for disordered gambling and whether they had lost a relationship or job of gambling.

Sociodemographic Measures

Data collected (categorisation in parentheses) included the following: gender (male, female), age, marital status (single, married/cohabiting, divorced/separated/widowed), education (no qualifications, GCSE/O-Level/A-Level, university/professional).

Procedure

Data was collected between March 2011 and March 2015 from clients who were voluntarily seeking treatment at the National Problem Gambling Clinic (NPGC). All clients entering the service during the study period completed the questionnaires prior to meeting with a clinician for their first appointment. The questionnaire included the PGSI and the clinical scales (PHQ-9, GAD-7, AUDIT-C) and the questions listed above. Demographic details were collected from a referral form completed by the clients prior to assessment. During the study period, 100% clients (1226 participants) consented to the use of their data. All participants were subsequently classified as problem gamblers (scoring 8 or more on the PGSI). All data was collected in accordance with NHS regulations of standard clinical practice.

Statistical Analysis

Once subjects had been identified, they were separated into two groups—no offending committed to finance gambling (non-offending group) (n = 705, 57.5%) and offending committed to finance gambling (offending group) (n = 521, 42.5%). The groups were then compared in the following areas—demographics, psychiatric and family history, and clinical scores (PGSI, PHQ, GAD-7 scores) and gambling history.

Standard descriptive statistics were produced in order to characterize our sample. Please note that not all categories total 100% because of missing data. Significant differences between our populations were determined using the Mann-Whitney U test (for ordinal variables) or χ2 test (for categorical variables) as appropriate. All analyses of data were carried out using Statistical Package for the Social Sciences (SPSS) 22.0. A significance level of 5% was adopted for all analyses.

Results

The overall prevalence of offending behaviour to finance gambling reported in those presenting to the NPGC was 42.5% (42.7% of males and 39.8% of females). There was no significant difference in age or gender between the group who had offended and those who had not. A greater proportion of the offending group was single or not married/cohabiting and had a lower level/no qualification compared with those who had not offended to finance gambling. The median salary earnings in the offending group were significantly less than the non-offending group: £6000+ a year less (see Table 1).

High rates of mental illness were found in both groups (Table 2). There was a statistically significant increase in a mental health diagnosis, a family history of mental health problems and previous gambling disorder treatment within the group who has offended to finance gambling. Moreover, a higher proportion of those in the offending group had experienced childhood abuse, specifically sexual abuse and multiple abuses compared with the non-offending group. A series of further Mann-Whitney U test showed clients who had offended to finance gambling scored significantly higher on the clinical measures (see Table 3); specifically, they had significantly higher median scores on the PGSI (indicating higher gambling severity) and the PHQ-9 and GAD-7 (indicating more severe depression and anxiety).

Table 2 Psychiatric and family history
Full size table
Table 3 Clinical scores
Full size table

Overall, those who reported offending behaviour to finance gambling started gambling at a significantly significant earlier age (by 2 years) and had gambled more days in the month before treatment started (Table 4). The offending group was more likely to have lost a job and a relationship due to gambling. Between groups, there was no significant difference between amount of current debt and the median total losses.

Table 4 Gambling history
Full size table

Discussion

This study adds to growing evidence demonstrating a relationship between disordered gambling and criminality. To our knowledge, this is the first study to compare sociodemographic, gambling behaviour specific and clinical variables in treatment-seeking gamblers within the UK who had offended to finance gambling compared with those who had not. Overall, 42.5% of participants had engaged in offending to finance gambling, largely consistent with overall offending behaviour the small amount of treatment-based international studies (e.g. Blaszczynski and McConaghy 1994; Meyer and Fabian 1992; Potenza et al. 2001). Furthermore, analysis indicates that a greater proportion of such offenders were single or not married/cohabiting and had a lower level/no qualification compared with the non-offenders. The median salary earnings in the offending group were significantly less than the non-offending group: £6000+ a year less.

High rates of mental illness were found in both groups (Table 2). There was a statistically significant increase in a mental health diagnosis, a family history of mental health problems and previous gambling disorder treatment within the offending group when compared with the non-offending group. Moreover, a higher proportion of those in the offending group had experienced childhood abuse, specifically sexual abuse and multiple abuse compared with the non-offending group. A series of further Mann-Whitney U test showed clients who reported offending behaviour scored significantly higher on the PGSI, PHQ-9 and GAD-7 indicating increased severity of gambling, depression and anxiety (see Table 3).

Overall, those who reported offending behaviour to finance gambling started gambling at a significantly significant earlier age (by 2 years) and had gambled more days in the month before treatment started (Table 4). This group of offenders were more likely to have lost a job and a relationship due to gambling. Between groups, there was no significant difference between amount of current debt and the median total losses.

The functional relationship between gambling and crime in multifaceted, and there is no consensus as to the link (Adolphe et al. 2019). A range of theoretical perspectives has been proposed, such as that crime is committed to subsidise gambling or gambling-related debt (Blaszczynski and McConaghy 1994), that antisocial personality characteristics and ultimately criminal behaviour surface due to the stress of gambling debt and the need for it to remain hidden from significant others (Blaszczynski and McConaghy 1991), that gambling can be considered an activity synchronous with other unlawful endeavours (Meyer and Stadler 1999) or that other factors may mediate the relationship such as a disinhibiting co-occurring substance use or impulsivity disorder (Preston et al. 2012). Notwithstanding the range of theoretical perspectives, as suggested by Perrone et al. (2013), May-Chahal et al. (2017) and Lahn and Grabosky (2003), the links can be differentiated by three principal explanations: First, that there is an instrumental link between crime and gambling, where, for example acquisitive crime may be used to get money to gamble or to pay off a gambling-related debt (Blaszczynski and McConaghy 1991). Second, the presence of a co-symptomatic link in that gambling and offending stem from the same underlying disorder (such as a comorbid disorder) (Blaszczynski and Silove 1996) and third a coincidental link (no relationship between the two) (Blaszczynski and McConaghy 1994).

Our findings correspond more closely to the first two links. Individuals explicitly admitted to committing crime to finance gambling, indicative of an instrumental link. Unmanageable debt has been shown to be a serious outcome of disordered gambling (Downs and Woolrych 2009, 2010) and can lead to a range of social harms that go beyond the individual including criminality (Langham et al. 2015). A major risk factor for gambling-related criminality has been shown to be excessive financial problems (Ledgerwood et al. 2007; Momper et al. 2010; Potenza et al. 2000). Once an individual who is in debt depletes all the legal opportunities to obtain money (e.g. salary, loans), they become anxious and depressed, and may become preoccupied with obtaining money that they seek alternative and potentially illegal means to obtain money as judgement is compromised (Rosenthal and Lesieur 1996).

The study established that there was a significant difference between those who had offended and those who had not; participants who report gambling-related offending had more severe psychopathology compared with clients who reported none (suggestive of a co-symptomatic link). A greater proportion of the offending group had a lifetime mental health diagnosis, had previous gambling disorder treatment and greater anxiety, depression and problem gambling scores. These results are in accordance with previous studies that have demonstrated that disordered gamblers with a history of arrest or criminal record show increased gambling-related problem symptomology (Ledgerwood et al. 2007; Mestre-Bach et al. 2018b; Potenza et al. 2000) and are more likely to have depressive symptoms (Momper et al. 2010) or anxiety (Ellis et al. 2018) and have received mental health treatment (Potenza et al. 2000). Such a finding suggests that crimes committed by disordered gamblers may be indicative of a more macro disturbance in the biopsychosocial functioning of those individuals. However, it must be noted that even though the relationship between gambling and crime in our sample is indicative of both an instrumental and co-symptomatic link, the exact framework integrating the two is still unspecified. A framework that can go some way to expounding the link is one that includes impaired impulse control. Pathological gambling has been shown to be associated with elevated impulsivity (Hodgins and Holub 2015) and elevated impulsivity is also associated with criminality (Lynam and Miller 2004). Such a framework would explain why not all individuals with a diagnosis of disordered gambling and related financial problems engage in criminal behaviour. The mediating effect of impulsivity on gambling and crime warrants further exploration.

As well as in a more severe psychopathology, the offending group also had higher rates of childhood abuse (multiple), a family history of mental health difficulties and gambling-related harms such as relationship and job loss. Previous studies have shown that individuals with gambling problems report greater childhood maltreatment and traumatic life events (e.g. Afifi et al. 2010; Hodgins et al. 2010; Roberts et al. 2017). Maltreatment has been shown to significantly increase the likelihood of criminality at a later stage and that the probability strengthens with the experience of several forms of maltreatment (Currie and Tekin 2012). In addition, a higher proportion of the offending group had lost a job or a relationship due to gambling and was more likely to be single/unmarried. In our group of participants, it is quite probable that the offending carried out may go beyond that explained merely by instrumental or co-symptomatic links but may be mediated by other multiple factors such as distinctive demography, impulsivity and varying clinical and personality traits (Granero et al. 2015), and highlights that criminal behaviour seldom appears in isolation (Petry et al. 2013). The finding also fits with the negative family and life experience in the ‘emotionally vulnerable problem’ pathway as described in the pathways model of gambling (Blaszczynski and Nower 2002).

The sample of gamblers who had committed crime to finance gambling also spent more days gambling and had an earlier age onset of gambling. Recent research in a gambling treatment-seeking sample revealed that those who started gambling at a younger age reported more severe gambling problems and were more likely to have committed an unreported crime (Sharman et al. 2019. In the pathways model (Blaszczynski and Nower 2002), the third pathway, the ‘antisocial impulsivist’ problem gambler is characterized by early age onset of gambling, early entry into offences related to gambling, poor interpersonal relationships and general psychopathology. Such a pathway may also go some way to explain our findings. It has been shown that these gamblers have poor treatment compliance rates and respond poorly to intervention (Blaszczynski and Nower 2002). Clinically, this pathway may contain different implications for selecting which treatment interventions and management strategies to carry out (Blaszczynski and Nower 2002).

Limitations

Participants were drawn from a treatment-seeking population and this figure may therefore not be an accurate reflection of those with a history of offending and disordered gambling in the general population. We did not corroborate self-reported offending history with official records (e.g. the Police National Computer Database); there is therefore the risk that offending to finance gambling was also underreported. Accordingly, previous research has shown that disordered gamblers are prone to underreporting as they do not always view their behaviour as illegal (Bergh and Kühlhorn 1994) or feel uncomfortable disclosing or fear reprisal (Adolphe et al. 2019). Moreover, we did not ask what type of crime the gamblers committed to finance gambling or whether other crimes had been committed not linked to gambling. Most crimes committed by problem gamblers are non-violent acquisitive crimes, such as fraud, forgery, embezzlement, larceny and tax evasion (Blaszczynski and Silove 1996); it would be useful for any future studies to collect this information.

Conclusion

Notwithstanding the limitations, this study adds to the increasing evidence that disordered gamblers who offend make up a distinct and complex subgroup and have distinct vulnerabilities. Such findings will be useful to clinicians involved in the assessment and management of problematic gambling. Our study established that within the offending group, illegal acts were linked to gambling severity. In addition, the sample who admitted criminality to finance gambling also had more severe psychopathology, early negative life experiences and added social stressors. Consequently, the gambling-crime relationship appears to be mediated by a diverse array of factors such as impulsivity and may be an indicator of a more acute presentation. As such, rather than targeted therapy at addressing disordered gambling or punitive measures of punishment for offending, there is an argument for a more holistic and collaborative approach that takes all related factors into consideration which could have a significant impact on the client. Clients with disordered gambling could be afforded the same opportunities.