Review of Self-exclusion from Gambling Venues as an Intervention for Problem Gambling

Self-exclusion programs are required to be provided by gambling operators in many international jurisdictions in an attempt to provide an option for those who have gambling problems to avoid further gambling. However, minimal robust and comprehensive research has been conducted to evaluate the effectiveness of self-exclusion programs. There is much scope for reform and greater cohesion between jurisdictions, particularly neighbouring jurisdictions that would offer greater protection to individuals and industry bodies. This review outlines the evidence surrounding existing self-exclusion strategies, the benefits and limitations of such programs, and provides potential recommendations for an effective intervention program. Research suggests that self-exclusion programs are under-utilised by problem gamblers and are not completely effective in preventing individuals from gambling in venues from which they have excluded, or on other forms. Nonetheless, self-report indicates that self-excluders generally experience benefits from programs, including decreased gambling and increased psychological wellbeing and overall functioning. There are many areas in which existing programs could be improved, such as providing more resources for excluded individuals and reducing barriers to program entry, and more research is needed. However, self-exclusion programs are an important component of any public health strategy that aims to minimise gambling-related harms and these should be based as far as possible on empirical evidence for effective program components.


Introduction
The extent to which government bodies regulate consumer behaviour is dictated in part by the most appropriate balance of individual freedom, personal choice and responsibility, and necessary safeguards and protection strategies to minimise potential harm. It is difficult for outside parties to verify appropriate limits for gambling behaviour in the absence of regulation and effective policies. In addition, appropriate gambling behaviour is extremely dependent on personal situations and factors and self-imposed attempts to limit gambling are often not effective for those who are most likely to exceed appropriate limits and gamble excessively. Although personal resolutions and willpower can have some use in controlling and modifying behaviour, there are no penalties for reversing these. This reduces the capacity for sustained control particularly given that problem gamblers, particularly electronic gaming machine (EGM) players, 1 often experience a loss of control and are more likely to exceed spending limits when they consume alcohol or are in certain emotional states, such as feeling bored, lonely, stressed or sad (Productivity Commission 2010).
In most societies internationally, the majority of adults gamble with annual participation rates ranging from 65 to 82 % (Fong et al. 2011;Petry et al. 2005; Productivity Commission 2010; Wardle et al. 2011;Wood and Williams 2010). Although most people gamble relatively infrequently and within affordable means, problem and the more severe pathological gambling are increasingly recognised as significant public health issues with a prevalence of 1-4 % in the adult population (Fong et al. 2011;Petry et al. 2005; Productivity Commission 2010; Wardle et al. 2011;Wood and Williams 2010). Problem gambling is characterised by difficulties in limiting money and/or time spent on gambling, leading to adverse consequences for the gambler, or others, such as health and psychological disorders, bankruptcy, or crime (National Research Council 1999;Neal et al. 2004). Problem gambling appears to be caused by a complex interaction between individual factors and a range of social and environmental influences (Blaszczynski and Nower 2002;Hodgins et al. 2011). Given the inherent risks involved with gambling, there is a continuum of risk of hard associated with this activity with the nature and severity of experienced harm being related to the frequency and level of expenditure (Currie et al. 2006(Currie et al. , 2012Schaffer 2005). Problem gambling is not a chronic condition and evidence from longitudinal studies indicates that people shift between levels of harm over time (Currie et al. 2012;LaPlante et al. 2008;Svensson and Romild 2011). Gamblers who are relatively free of any symptomatology are referred to as low-risk gamblers, and moderate-risk gamblers are those experiencing some related difficulties, such as gambling more than they intended, but without significant impairments to other areas of their lives (Currie et al. 2006(Currie et al. , 2012. Given the difficulties in problem gamblers, by themselves, to effectively gambling within pre-set limits (Lalande and Ladouceur 2011), there are grounds for governments to ensure that resources are available to assist players. Although industry bodies may selfimpose the responsibility to provide appropriate harm minimisation interventions, government regulation provides the necessary power to ensure that these are effective, empower venues to enforce their commitments and impose penalties for industry operators and individuals who do not comply with agreed strategies.
Self-exclusion is an extreme form of pre-commitment, in which gamblers who believe that they have a problem can voluntarily bar themselves from entering one or more gambling venues to prevent them from gambling. Under a formal self-exclusion program, the individual agrees that the venue and/or government regulators authorise gaming staff to deny them access to the venue. This agreement places the responsibility on the individual as they risk removal for breaches and can possibly be charged with trespass. Nominated periods of self-exclusion vary from 6 months to irrevocable lifetime bans, and detected breaches may incur additional penalties beyond simply being removed from venues. Evolving from informal banning procedures used by casinos to evict problematic or unruly patrons, self-exclusion programs have become the predominant harm-reduction strategy used by the gaming industry (Nower and Blaszczynski 2008). Such programs are designed to limit access to gaming opportunities and provide problem gamblers help to cease or limit their gambling behaviour (Blaszczynski et al. 2007). Self-exclusion agreements have historically been industry-driven, but an increasing number of jurisdictions are introducing legislation requiring the provision of programs of self-exclusion reflecting increasing social and community concern in regards to problem gambling and the availability of gambling.
Despite the implementation of self-exclusion strategies in numerous international jurisdictions, minimal research has been conducted to evaluate the effectiveness of selfexclusion programs. Furthermore, policies differ between jurisdictions making comparisons and generalisation of results difficult. Therefore, there is little information to inform best practice and what elements should be included in programs. However, although the evidence is not comprehensive, preliminary findings indicate that this type of precommitment arrangement has significant benefits for problem gamblers.
Creating an effective self-exclusion program is a challenge as self-exclusion agreements currently suffer from various limitations that reduce both the extent to which they can be enforced and success in preventing individuals from gambling. The deficiencies of selfexclusion strategies have been recognised and efforts are being taken to implement changes that offer potential solutions to some of the limitations. There is much scope for reform and greater cohesion between jurisdictions, particularly neighbouring jurisdictions that would offer greater protection to individuals and industry bodies. This review will outline the evidence surrounding existing self-exclusion strategies, the benefits and limitations of such programs and potential recommendations for an effective intervention program.

Features and Foundation Principles of Self-exclusion Programs
The features and principles of a self-exclusion program should be fully understood by individuals who wish to self-exclude, employees of gaming venues, gaming venue operators and regulatory bodies. This is essential in order to clarify expectations regarding the role and limits of all parties including legal and governmental authorities and avoid unrealistic expectations and unfair criticisms.
Self-exclusion agreements generally do not constitute a formal contract enforceable by law (Napolitano 2003). By signing a self-exclusion agreement, individuals typically agree to certain obligations and forgo some rights under an agreement with an operator that is offered voluntarily or enforced by law. These include: • for nominated venues, they agree not to enter gaming areas, not to play gaming machines or not to enter the venue at all • authorising staff to stop them from entering or remaining in a gaming area or venue that they are excluded from • authorising for photographs and personal details to be taken and disseminated to relevant venues and for venues to display the photographs • waiving the right to sue nominated venues, their staff or the program administrator on the grounds of assault, defamation or failing in a duty of care to exclude • accept their personal responsibility to stay away • acknowledge that nominated venues or their staff have no legal duty implied by the self-exclusion deed.
Most venues advertise self-exclusion programs on their websites and through pamphlets and displays in venues. Individuals wishing to self-exclude can usually register at a venue, sign the agreement and have their photograph taken. They are typically advised that there is help for problem gambling, receive information about the self-exclusion program and the nature of the agreement, including their responsibility to uphold it and may be provided with referral information for further help. Self-exclusion agreements are typically offered for periods of 6 months, 1-5 years or as lifetime bans. Venue security personnel typically enforce self-exclusion policies. A breach is recorded if the person is discovered contravening their agreement. The first time a person is discovered breaching self-exclusion, they are typically asked to leave the venue. In some jurisdictions, this process involves the attendance of a law enforcement officer or representative from the gambling regulator. While rarely used, someone breaching a selfexclusion agreement can be charged with an offence and/or fined.

Literature Review: The Effectiveness of Current Self-exclusion Programs
This review describes the findings of the relevant studies conducted on self-exclusion programs available in the published academic and grey literature. It is intended to provide a comprehensive understanding of the available evidence to date that is relevant to the establishment and implementation of a self-exclusion program. This review is limited due to the few comprehensive evaluative studies that have been conducted for self-exclusion programs. Furthermore, the studies included have methodological limitations that reduce the extent to which results can be used to improve existing programs or inform new strategies and older studies have been excluded for lack of relevance. Table 1 outlines the details of the major studies that have been conducted in regards to self-exclusion programs.

Quebec
Self-exclusion programs are offered in each of the casinos in Quebec and range from periods of 3 months to 5 years. The program is run by the casino security department and is publicized in pamphlets available in the casino. To register, the individual is taken to a private office where they complete and sign a consent form and a photograph is taken. In the casinos, security agents are trained in the identification of self-excluded individuals and if a self-excluded individual is identified they will be approached and ask to leave.
The evaluations completed indicate that between 73 and 95 % of self-excluders are probable pathological gamblers (Ladouceur et al. 2000. The most popular period of self-exclusion appears to be 12 months (46-66 %), as compared to 2 years (21-25 %), or 6 months (9-33 %). The majority of participants report that they decided to self-exclude themselves (74-88 %), although family and friends also appear to play a role in the decision to self-exclude. Financial problems constituted the main reason for self-exclusion and the majority reported that they were not able to stop gambling of their own accord.
The self-exclusion program has been shown to be linked to a reduction of pathological gambling habits and gambling-related problems , although direct causality cannot be determined. The levels of reported satisfaction with the program was generally high (80 %) amongst participants and the majority (77 %) indicated that they would use self-exclusion programs again (Ladouceur et al. 2000. At the beginning of the program the majority of participants believed that the self-exclusion program would be effective (62-97 %) in helping them stay away from casinos. However, 70 % of selfexcluders reported gambling in casinos or elsewhere during their ban and 11-55 % of gamblers broke their contract. Those who returned did so an average of six times. High dropout rates amongst participants in the study may mean that these figures underestimate the proportion of individuals who broke their agreements. Criticism also existed regarding the self-exclusion program and many gamblers felt that the programs did not provide them with sufficient resources on problem gambling treatment and support during the ban period, that the detection process was weak, the program was not well advertised and they should be able to renew a self-exclusion agreement without going back to the casino (Ladouceur et al. 2000).
A third study was undertaken in Quebec following modifications made to the selfexclusion program in 2005 (Tremblay et al. 2008). In the new procedure, individuals have the opportunity to meet with a self-exclusion counsellor at the beginning of the selfexclusion period. The counsellor is a psychologist, independent from the casino and located outside the casino. Additionally, telephone support from the counsellor is available to direct the self-excluder toward appropriate resources during the ban period. Finally, a counselling meeting is required at the end of the self-exclusion period to evaluate the situation and provide appropriate information about chance and responsible gambling before re-entry to the casino is permitted. If individuals refused to attend the mandatory meetings they are permitted to sign a regular agreement and given an information sheet to explain the self-exclusion service with an option to contact a counsellor if desired.
Three-quarters (n = 857) of all individuals who signed a self-exclusion agreement between November 2005 and May 2007 opted for the improved program (Tremblay et al. 2008). Although 40 % of the self-excluders requested an evaluation, only one-third came to the initial meeting. This represents 15 % of self-excluded gamblers and is slightly higher than the 10 % of self-excluders in a regular self-exclusion program who report seeing a counsellor (Ladouceur et al. 2000). Very few (n = 5) participants initiated calls to their counsellor for support during the self-exclusion program. Although 74 % of participants reported in their initial meeting that they were very or totally able to resist the impulse to gamble, 46 % of participants who responded to the follow-up survey reported breaching their self-exclusion contract. However, the majority of participants (82 %) reported that the self-exclusion program was 'very' or 'totally' effective. Analysis of responses at the initial and final meeting showed that there was a significant reduction in time and money spent gambling as well as the intensity of negative consequences of gambling. Participants appeared to have fewer symptoms of pathological gambling as well as fewer symptoms of depression, anxiety and at-risk alcohol consumption.
These studies include small, non-representative samples of gamblers who have entered self-exclusion agreements in Quebec and as such the results are limited in the extent to which conclusions can be extrapolated. However, the general findings indicate that although self-exclusion programs are not highly effective in preventing individuals from gambling, they do appear to be associated with a reduction in gambling behaviour and problem gambling severity.

Nova Scotia
An evaluation of a trial of a self-exclusion program implemented in 45 EGM retail sites in Nova Scotia in 2004 was conducted with gaming venue employees, regular EGM players, and confederate players to test the detection and enforcement policies (Schrans et al. 2004). The study did not examine the impacts of the self-exclusion program on problem gamblers, but tested the compliance of gaming venue employees. It was tested in a limited area comprised of rural and small urban communities so it is not possible to extrapolate how the program would impact large urban communities. As it only included a small number of venues and 36 ''self-excluded'' player participants it is also much smaller than any actual self-exclusion program. The actual program implemented is not described in detail, but appears to involve a centralised self-exclusion database with photographs and participant information circulated to all participating venues. Gaming venue employees were responsible for identifying self-excluded players and filing reports.
The evaluation (Schrans et al. 2004) found that following program implementation venue identification rates and compliance with program protocols declined significantly. After only 3 months identification rates fell to one-third of detection rates at the start of the trial and breach submission reports dropped by about half. The overall identification rate for play sessions by program participants was 23 %, meaning that 77 % of play visits were not detected and/or correctly reported upon. This is in spite of monthly formal training sessions and notification packages sent to venues. More concerning was the poor detection rate, even with only 28 players on the ''excluded'' list in the 1st month. Only one in every three play visits was being detected for regular players and one in ten for unfamiliar players. Furthermore, only 42 % of reports filed were accurate in identifying a ''true positive'' play visit demonstrating high levels of mistakenly identifying player breaches and unnecessary paperwork. Also concerning was the finding that one-third of local players taking part in the study encountered some issues with breaches in confidentiality. Most of these were not malicious or deliberate disclosure but rather unintentional or careless breaches. The report concluded that the process test found that the retail monitoring component of the program proposed for multi-site EGMs was not sufficient to support the programs objectives or expectations. The results indicated that it was not appropriate or reasonable to rely on gaming venue employees to subjectively detect and accurately report on self-excluded players. Finally, the authors concluded that changes identified to improve the program are likely to be too cumbersome, expensive and impractical to be implemented and, moreover, are unlikely to assure the required improvements in venue performance.

Canada
A review of the use of self-exclusion programs for casinos in Canada was conducted by Williams et al. (2007). This review estimated that, based on self-exclusion data from 2005 for the seven Canadian provinces with casinos, between 0.6 and 7.0 % of problem gamblers signed up to self-exclude. These fairly low utilisation rates suggest that programs need to be promoted more effectively and potentially modified to make them more attractive as a suitable strategy to control gambling for problem gamblers.
A telephone interview was conducted with 300 randomly selected self-excluded individuals from seven Canadian provinces to evaluate the effectiveness of self-exclusion programs (Verlik 2008). The majority (49 %) were 35-54 years old, participants were most likely to be male (58 %), employed full-time (68 %), and play EGMs (57 %). The majority of participants (68 %) appeared to be high-risk problem gamblers and 17 % were at moderate risk, although only one-third had ever sought formal help for their gambling problems. Participants reported finding information about the program through family or friends (42 %), staff at gaming venues (24 %) or posters, pamphlets and brochures (20 %). When asked about the factors that were important in their decision to enter the program, the majority of participants endorsed the desire to stop gambling altogether as being extremely or very important (81 %) and the threat of personal consequences (70 %) and other people (65 %) were also viewed as very important factors.
Most participants were happy with the information provided about the self-exclusion program. Participants rated casino staff (74 %), the Alberta Gaming and Liquor Commission website (84 %), and printed information (73 %) as very or somewhat effective in providing them with information (Verlik 2008). However, there was a clear demonstration of the need for more staff training as only 69 % of self-excluders rated staff who enrolled them in the program as being very or somewhat effective in explaining the program and the responsibilities of each party. Similarly, only 68 % of participants rated the program's current enrolment and registration process as very or somewhat effective indicating room for modifications and improvements. Of those who felt that the enrolment periods were ineffective, 41 % stated that they would have liked to have the option of a lifetime ban indicating the importance of having options for time frames of self-exclusion periods. Although most participants appeared to be satisfied with the level of information they were provided with, only 48 % felt that they were completely or very informed about the penalties for breaking the program agreement. Participants were mostly (51 %) satisfied with the severity of penalties and did not agree with an option to remove self-exclusion bans before the time period had expired (72 %). Overall 67 % of respondents rated the overall effectiveness of self-exclusion programs as somewhat or very effective. Just over half of the self-excluded participants in Canada admitted to breaching their agreement, and those that breached did so frequently (Verlik 2008). Of participants that breached their agreement, only 48 % were recognised and 81 % stated that it was very easy to gain entry to excluded venues. At least 50 % of participants currently enrolled in selfexclusion programs stated that they planned to re-enrol in the program, with 74 % planning to re-enrol for three years. The majority of participants (68 %) stated that facial recognition technology and mandatory identification checks would be very effective, as would mandatory identification checks for jackpots with winnings forfeited (61 %) and more information being provided about problem gambling services (51 %). The options of lifetime enrolment (47 %), a mandatory re-entry program (45 %), and third party initiation (30 %) were not endorsed by as many self-excluded participants and 30 % of participants also viewed third party initiation of self-exclusion program to be somewhat or very ineffective.
Focus groups were conducted with 76 individuals with self-exclusion program experience across seven Canadian provinces (Responsible Gambling Council 2008). This sample was not intended to be representative of all Canadian self-excluders and as such results must be interpreted with caution. Participants reported that following self-exclusion their gambling behaviour reduced in terms of number of sessions, and time and money spent gambling. One-third reported breaching their self-exclusion agreements; among these 70 % were not detected and of those who were detected 63 % went on to breach again suggesting that being detected was not a deterrent for the majority of participants. The majority of participants (59 %) reported that they had engaged in other forms of gambling, not covered by their self-exclusion agreements, during their bans. Of the focus group participants, only 30 % reported abstaining from all gambling activities during their ban.
Results of the focus groups indicated that self-exclusion programs are an important tool for patrons dealing with gambling problems (Responsible Gambling Council 2008). Many participants reported that self-exclusion agreements played a significant role in helping them to stop gambling and how good it felt to be in control of their gambling. Even those who were not successful in quitting entirely reported reductions in amounts of time and money spent and the frequency of gambling after they had self-excluded. The participants reported that staff should be better trained and self-exclusion should be dealt with in a more compassionate and supportive manner with resources and options provided for selfexcluders to assist in controlling gambling. Although there was mixed support for the use of player cards or identification checks when entering venues, there was a consensus that bans need to be better enforced. Participants acknowledged their individual responsibility with regards to self-exclusion, but stated that bans need to be taken more seriously by venues and consequences for breaching bans are not severe enough. Participants reported that bans should be used in more venues, including EGM sites, not just casinos and they should be promoted more widely. There was general scepticism in the extent to which gaming venues wanted to have robust self-exclusion programs because these might negatively affect their business and participants felt that a third party should regulate any program, including penalizing venues that do not comply.

Missouri
The Missouri Gaming Commission has provided researchers access to the censored roster of enrolees in the Missouri Voluntary Exclusion Program (MVEP) to investigate the long-term effectiveness of the program in helping participants change their gambling behaviour. The MVEP was created in 1996 as the first state-wide self-exclusion program in the US. Participants volunteered to ban themselves for life and assumed responsibility for not entering any Missouri casino. The Missouri Gaming Commission ensures that all participants' names are removed from marketing lists that self-excluded individuals are prohibited from cashing cheques in casinos and requires identification before compensating any jackpot winner of $1,200 or more. If self-excluded individuals are caught entering a casino they may be arrested and charged with trespassing (Missouri Gaming Commission 2008). Nower and Blaszczynski (2006) examined the characteristics of gamblers enrolled in the MVEP between 2001 and 2003 and analysed the data based on gender differences. The gender ratio for self-excluded gamblers was found to be approximately equal and a high percentage of both genders endorsed EGM play. Female self-excluders were more likely than males to be older at the time of application, African American, and either retired, unemployed or otherwise outside the traditional workforce. In addition, female self-excluders were more likely to report a later age of gambling onset, a shorter period between onset and selfexclusion, a preference for non-strategic games and prior bankruptcy.
A subsequent analysis of a subset of individuals enrolled in the MVEP between 2001 and 2003 and analysed the data based on age (Nower and Blaszczynski 2008). Compared to younger or middle-aged adults, older adults were more likely to have gambled longer before self-exclusion, to be married and/or retired or unemployed, and to express a strong preference for nonstrategic forms of gambling. In addition they were nearly four times as likely to self-exclude in an effort to prevent suicide. The authors concluded that older gamblers represent a distinct subgroup of problem gamblers whose gambling behaviour is likely tied to situational factors that prompt initiation and rapid escalation of gambling.
Reasons for registering for self-exclusion across all groups included gaining control, needing help, and hitting rock bottom Blaszczynski 2006, 2008). The findings of these studies are quite preliminary and are limited by the lack of valid screening instruments, use of categorical variables and extent to which results can be extrapolated to other jurisdictions given the lifetime condition of the ban, which is not the typical exclusion period.
In 2007 researchers attempted to contact a randomised, representative sample of participants in the MVEP to conduct a telephone survey (Nelson et al. 2010). This represents one of the first long-term follow-up studies of a self-exclusion program. The time between MVEP enrolment and follow-up interview ranged from 3.8 to 10.5 years with an average of 6.1 years (SD = 1.6). After entering the program 25 % of participants reported quitting all gambling, 18 % reported quitting casino gambling and the remainder (58 %) reported not quitting any gambling. Of those surveyed only 13 % had not gambled at all since enrolling in MVEP. Among those who reported quitting all types of gambling, 46 % gambled at some point after MVEP enrolment. Over half (60 %) of participants surveyed gambled in the 6 months prior to the interview with lowest recent participation amongst those who attempted to quit all forms of gambling. Prior to entering the MVEP 97 % of participants gambled in Missouri casinos and after self-exclusion only 8 % reported gambling in Missouri casinos, although 74 % went to casinos in other jurisdictions. Following enrolment, 16 % of participants attempted to enter Missouri casinos an average of 4.7 times, of these, only 50 % were caught indicating a lack of strict policy enforcement. The majority (81 %) of participants who continued to gamble regularly after enrolling reported gambling less than before and none reported gambling more.
When asked what promoted enrolment in the MVEP, 23 % of participants had stated that others had influenced them, including supportive, coercive and contagion influences (Nelson et al. 2010). The majority (77 %) provided self-related reasons, most commonly financial worries as well as an inability to control gambling and recognition of having a problem. Most participants (68 %) expressed satisfaction with the program, however nearly one-third expressed dissatisfaction including complaints regarding the permanence of the ban and inadequate explanations at registration. Overall participant's gambling severity scores decreased significantly following MVEP enrolment, although this was reported retrospectively, and there was slight reported improvement in quality of life. Almost 60 % of participants reported receiving some form of treatment or self-help and participants were more likely to report involvement in therapy after enrolment (34 %) than before (15 %). Treatment and self-help involvement was significantly related to post-MVEP quality of life as well as gambling abstinence.

Australia
It is estimated that there are around 15,000 self-exclusion agreements in force in Australia (Productivity Commission 2010). This suggests that between 9 and 17 % of problem gamblers in Australia are currently self-excluded (Productivity Commission 2010). State gambling prevalence surveys have reported that 31-61 % of problem gamblers surveyed have attempted to self-exclude (AC Nielson 2007; Queensland Government 2006Government , 2007 South Australian Department of Health 2006). However, these surveys are often limited by the small, non-representative samples included and methodological issues that may overestimate the number of self-exclusion agreements (e.g., by asking individuals about their attempt to self-exclude as opposed to actual self-exclusion agreements entered into). A survey of problem gamblers undergoing counselling found that 39 % of problem gamblers had self-excluded from gaming venues (Productivity Commission 2010). However, treatment-seeking gamblers are more likely to enter (or be required to enter) selfexclusion agreements than problem gamblers who do not seek formal treatment. In a review of evidence available the Productivity Commission (2010) concluded that approximately 10-30 % of problem gamblers have current self-exclusion agreements in place, which is higher than figures reported from Canada by Williams et al. (2007). A pilot study of 135 self-excluded gamblers was conducted in New South Wales, Australia between 2003 and 2005 ). This survey revealed that 79 % of male and 80 % of female participants gambled on EGMs while on a self-exclusion agreement and 45 % of male and 33 % of female participants gambled at a venue from which they were specifically excluded. Those who breached their agreements typically did so on at least 10 occasions. On 56 % of occasions self-excluded men were identified by venue staff and 71 % of women were identified. About one in three broke their contract within a month of the agreement, and of those who breached contracts this was usually between one and 6 months after signing the agreement. The majority (75 %) started gambling within 6 months of signing the agreement.
Although the program appears to have limited effectiveness in stopping participants from gambling, the overwhelming majority of participants strongly supported the program . Participants stated that the program had been very helpful in regaining control of their financial affairs and overcoming relationship problems. Furthermore, many participants found the process of enrolling into the program empowering and saw it as the start of their recovery, which may be related to the skill of the counsellor at their initial interview. Finally, about 70 % of participants more than halved the amount they spent on gambling, indicating that the program had a positive impact on gambling expenditure.
A large-scale, multi-phase study of the motivators and barriers for help-seeking among gamblers was completed by the Centre for Gambling Education & Research (Hing et al. 2011). Analyses were conducted of responses from a combined sample of regular gamblers recruited via a national telephone survey and in culturally-diverse gambling venues, gamblers calling a telephone helpline and an online survey of problem gamblers in treatment. In the combined sample (N = 730), there were 153 non-problem gamblers, 82 low risk gamblers, 117 moderate risk gamblers and 346 problem gamblers (32 missing cases). Awareness of professional sources of help was very low, only 31 % of participants were aware of venue assistance, including self-exclusion programs, which was the second most known source of help, behind telephone counselling. Awareness of self-exclusion was highest among problem gamblers (36 %) and low risk gamblers (35 %), as compared to moderate risk gamblers (25 %) and non-problem gamblers (25 %). Younger gamblers (18-39) were more aware (37.5 %) that venues can offer assistance for a gambling problem than were older gamblers (28.2 %). When asked about factors that would motivate them to seek help for gambling problems, ''concerns from the venue where you were gambling'' was rated as the least important factor by the overall sample and among problem gamblers.
A study of self-excluders in Victoria was conducted to further the understanding of motivators and barriers to self-exclusion specific to hotels and clubs (Abbott et al. 2011). Interviews were conducted with 60 self-excluders (66 % female, mean age = 53 years, age range 28-86) who on average had 3 years' experience with self-exclusion (range = 0-10 years). Almost half of the participants (45.5 %) stated that it took them less than a month to self-exclude after finding out about the program. A further 15 % joined within 4-6 weeks, 9 % within 3 months, 7 % waited 6 months and 7 % waited a year before commencing self-exclusion. The majority of participants (73 %) stated that they felt very comfortable or enthusiastic about joining the program, and a further 20 % stated they felt reasonably comfortable or enthusiastic about this. It is possible that the retrospective nature of these accounts has moderated these responses given that anecdotal evidence indicates that many individuals are relatively distressed and anxious during the initiation process and research confirms that this option is often a final resort based on a crisis situation. Therefore, these responses are more likely to be an indicator of current levels of comfort and enthusiasm about the program.
Of the 24 motivators discussed, all were endorsed as being reasonably important by the majority of the participants (78 %). Responses are consistent with previous research and indicate that motivating factors include: emotional distress, financial crisis, a loss of control, and embarrassment related to being asked to leave. The least important factors, still rated on average as 'a lot' important included the immediacy of exclusion compared to counselling, physical health concerns, and approval from family members. Men were more likely to endorse the lack of time commitment, compared to counselling, as being important than women.
The most significant barriers to joining a self-exclusion program appeared to be admitting to oneself that they had lost control and needed external assistance to stop gambling. The majority of participants (85 %) also indicated that the fact that they did not want to stop gambling was reasonably important and 92 % indicated that they did not know what they would do with their spare time if they were not gambling and that they did not want others to find out about their gambling.
This research had a small sample that was self-selected and may not be representative of all self-excluders. The questions were somewhat awkwardly worded and many of the factors provided as motivators and barriers concentrated on a relatively small number of factors. Nonetheless, the research confirms previous studies that indicate emotional and financial distress and related feelings of being out of control are important motivating factors to join self-exclusion programs. The most important barriers to joining a selfexclusion program appear to be related to a desire to address gambling-related problems unaided and difficulty admitting that one has a gambling problem and needs assistance. Subsequently, efforts are needed to reduce the stigma related to problem gambling and awareness that it is a relatively common problem that does not imply a personal deficit. For example, campaigns can work to put a different face on problem gamblers by depicting a wide range of individuals who have gambling problems (to get away from the stereotypes) and the courage and strength it takes to admit to needing help. Addressing the shame and embarrassment related to admitting to needing help is essential to increase participation in self-exclusion programs.

New Zealand
In New Zealand, problem gambling is encompassed in a public health approach and selfexclusion is legislated by the Gambling Act 2003, which takes a product safety approach (Townshend 2007). An individual can exclude themselves at a gaming venue or by mail and agreements can last for any nominated period up to 2 years and only apply to the gambling area of the venue. If a self-excluded gambler breaches a ban they commit an offence and can be fined up to $500, the venue that allows them to gamble also commits an offence and can be fined up to $10,000 and may not be able to renew their gambling licence. The strict consequences of breaches have made gaming operators very wary of allowing self-excluded gamblers into their venues (Townshend 2007  Townshend (2007) conducted a small follow-up survey of 35 self-excluders from a single community problem gambling treatment survey. The majority (60 %) were male, all played EGMs primarily, 29 % had co-morbid substance use disorders and 20 % had other mental health disorders. A comparison of the information gathered and assessment and the follow-up survey found that there was a significant reduction in problem gambling symptoms and severity as well as money lost. Participants also reported greater control over their gambling and approximately 80 % stated they were abstinent from gambling, although two participants stated that their gambling remained unchanged. This was a small scale study of a non-representative sample and it is not possible to determine the impact of self-exclusion agreements separate from the effects of treatment. Furthermore, the participants had initiated self-exclusion agreements with the help of the service. However, it supports the use of self-exclusion agreements in combination with treatment for problem gambling.

Europe
Self-exclusion programs appear to be more effectively in European countries than in North American and Australia. In many European jurisdictions, individuals are required to show personal identification before entering a casino. For example, at casinos operated by Holland Casino, a computer system registers all visits and immediately identifies anyone who has requested a ban or visit limitation (De Bruin et al. 2001). The Netherlands also appears to have relatively high self-exclusion program utilisation rates with an estimated 25,000 agreements arranged between 1990 and 2002 (Nowatzki and Williams 2002). Amongst a random sample of 50 problem gamblers, 40 % had been reached by Holland Casino's prevention policy by either asking for a protective measure or being approached by Holland Casino about their gambling behaviour (De Bruin et al. 2001). Furthermore, awareness rates of visit limitation and self-exclusion policies appear to be relatively high (above 75 %) in the Netherlands amongst casino visitors and problem gamblers. However, even with this system a large proportion of self-excluded individuals eventually returned to the casino following the period of restriction. Of these some had a sharp increase in visiting frequencies in the following 6 months, although the majority of individuals eventually stabilised their visiting frequency at less than eight visits per month. Additionally, about half of self-excluded patrons reportedly found alternative ways to gamble during their selfexcluded period such as illegal gambling or EGMs outside of casinos (De Bruin et al. 2001). It is important to note that these statistics are over 10 years old and it is uncertain whether the existing program has a similar impact.
In a study of 450 regular gamblers (De Bruin et al. 2006), 83 % of the Holland Casino problem gamblers who were interviewed were aware of the entry limitations and visitation bans, 13 % had a form of entry limitation (two-thirds of this group were satisfied with this measure and 50 % had gambled elsewhere when the casino ban was in place). Of those problem gamblers that used amusement arcades, 59 % were aware of the exclusion lists, and 20 % had been placed on these lists at some point (80 % were satisfied with this measure and 75 % had gambled elsewhere during the ban).
The majority of participants who restricted their play went to gamble at another location during the restricted period. Slot machine gamblers from the casinos typically gambled in an amusement arcade, but also went to other establishments or gambled abroad. None of the banned arcade gamblers entered Holland Casino during their bans. Casino gamblers often started gambling aboard or in an illegal casino (Goudriaan et al. 2009). A more recent study would likely find that a substantial proportion of banned or restricted gamblers may gamble at online sites when they are restricted from gambling venues. Although the majority of respondents were satisfied with the restriction programs, a large proportion indicated that the effectiveness would be increased if the measures applied to both casinos and arcades. A similar detection system operates in Switzerland, entry is controlled and visitors are required to show a passport or identification card and all self-excluded individuals are registered in an electronic database (Haefeli 2005). In 2004 there were 3,396 bans in existence and 326 self-exclusion bans that had been lifted, which represented one gambling ban per 1,425 visitor entries. The majority of self-excluders were men (80 %) and EGM gamblers (54 %), and were most likely to be aged 31-40 (35 %), although one quarter were aged 18-30 and over one-fifth (22 %) were aged 41-50 (Haefeli 2008). Only 44 % of bans were for Swiss gamblers as the majority of self-excluded players were of other nationalities. The 19 casinos are all networked so gambling bans are applied throughout Switzerland and are monitored by a regulator. However, it is not known whether selfexcluded players gamble in other countries or venues that are often very close to the Swiss border, or participate in other forms of gambling.
An evaluation of self-exclusion programs in Europe gathered data from casinos in Austria, Germany, Switzerland and two Internet gambling sites (Hayer and Meyer 2011a, b). Participants were given a survey and up to three follow-up questionnaires and one interview. The casino sample included 152 people, the majority (72 %) of which were males (mean age 41.3), primarily played EGMs (49 %) and met criteria for pathological gambling (51 %) (Hayer and Meyer 2011a). Almost one-quarter (24 %) meeting criteria for problem gambling and one-quarter were non-problem gamblers. Participants appears to have a relatively long period of consideration prior to deciding to self-exclude although they reported that being able to stop gambling was very important and on average were relatively confident that they were able to succeed. The most common reason given for self-exclusion was having lost too much money (76 %) and the agreement was also deemed important as a preventive measure (60 %) and due to a loss of control (53 %). One-fifth of the participants reported seeking some form of help (including self-help and debt advice) in the past, although 45 % reported that they intended to do so in the future. These numbers may be inflated by multiple forms of help seeking indicated by a single respondent.
At 1, 6 and 12-month follow-up, the number of respondents markedly declined reducing the extent to which conclusions can be drawn from the results (Hayer and Meyer 2011a). However, the numbers of pathological gamblers markedly declined and fewer respondents reported being problem gamblers with a corresponding increase in the number of nonproblem gamblers in the sample. At one and 6-month follow-ups the majority of participants reported gambling less often and for shorter periods of time with lower stakes since exclusion. At the 12-month follow-up there appeared to be some rebound with approximately one-third of participants reporting no change in their gambling behaviour and up to one-quarter reporting increased gambling in the past 6 months. Participants appeared to experience a reduction in their urges to gamble, emotional stress and loss in quality of life across time periods. Feelings of control over gambling appeared to increase at 1 and 6 month follow-ups, but declined to almost the same level as at exclusion after 12 months.
The sample gathered from European-based Internet gambling sites (N = 259) was still primarily male (69 %), slightly younger (mean age 36.2) than the casino participants and included more problem gamblers (68 %) and non-problem gamblers (32 %) with no pathological gamblers (Hayer and Meyer 2011b). The decision to self-exclude appeared to be rather spontaneous with a short period of consideration and being able to stop gambling immediately was still relatively important, but to a lesser extent than for the casino sample. The reported confidence in being able to succeed with self-exclusion was similar to that reported by the casino sample. Participants indicated that they had self-excluded as a preventive measure (63 %), as they had lost too much money (52 %) and spent too much time on the site (36 %).
Similarly to the casino sample, the number of participants markedly declined at each time period for the Internet sample, with only nine participants reporting at the 12-month follow-up, again reducing the generalisability of results (Hayer and Meyer 2011b). Significantly fewer participants appeared to be problem gamblers at each time-period and approximately three-quarters of each sample reported gambling less frequently and twothirds gambled with lower stake sizes across time periods. At one and 6 months after selfexclusion over half of the participants reported spending less time gambling, and after 12-months, two-thirds reported no change in the past 6 months, indicating they had remained at shorter gambling periods since self-exclusion.
The authors concluded that individuals mostly make use of self-exclusion programs when negative gambling-related consequences are already evident (Hayer and Meyer 2011a, b). The proportion of pathological gamblers in both samples was lower than that found in samples of self-excluders in Canada (e.g., Ladouceur et al. 2000), but the level of help seeking behaviour appeared similar as to that found in other countries. Similarly to results from other studies, self-exclusion appears to have at least a short-term positive effect on gambling behaviour, problem gambling severity and psycho-social functioning, however the benefits may decline over time. The conclusions are limited by the small, nonrepresentative sample that declined over the length of the study, again reinforcing the importance of comprehensive evaluative studies.

Discussion
The assessments of self-exclusion programs internationally generally find that the majority of participants benefit from such schemes. These benefits include participants reporting decreases in gambling expenditure and improved financial circumstances; decreases in gambling frequency and time spent gambling; reduction in problem gambling severity and negative consequences of gambling; reduction in related psychological difficulties including depression and anxiety; and that they feel they have more control of their circumstances. Even without enforcement, self-exclusions may be somewhat effective because they allow problem gamblers to make a public commitment to stop gambling. Some problem gamblers will wish to avoid the potential embarrassment of being caught in a break of a self-exclusion agreement.
Although there is some evidence of the positive impacts associated with self-exclusion programs, there is little to indicate the magnitude of the causal link. A gambler's willingness to address their adverse gambling behaviours precedes self-exclusion in most cases. It is likely that this willingness, as well as the self-exclusion and potential referrals, results in better outcomes for the problem gambler.
Although patrons in self-exclusion programs have reported positive benefits, the current programs are in need of improvements to improve utilisation rates and outcomes over time. Current programs have been criticised and it has been argued that the industry have offered marginal support for programs, which thereby are falling short of individual and community hopes for effectiveness (O'Neil et al. 2003;Nowatzki and Williams 2002).
A key deficit in current self-exclusion programs is that the majority of problem gamblers do not enter into these agreements. There is a clear need to reduce some barriers to self-exclusion, such as limiting embarrassment in instigating process, which can be heightened in smaller rural communities where privacy is difficult to maintain. Similarly, it is important to remove any unnecessary complexities in the application and registration process, including for those who have limited proficiencies in English and unnecessary legal jargon. Individuals should be able to exclude from multiple venues in one step and have the ability to enacted agreements away from gaming venues, such as at a central administrative office, with a health or mental health treatment provider or legal professional, or via the Internet or mail.
Self-exclusion programs are partially dependent upon the ability of gaming operators to accurately identify program participants in order to detect and report violations of the selfexclusion agreement. The studies completed thus far indicate that the principal points at which self-exclusion programs are breaking down are the point of detection and the point of enforcement of expectations as expressed in the act of self-exclusion . Identifying self-excluded patrons from photographic information is highly problematic from the venue's perspective and the problem of detection is compounded with growth and expansion of programs. Evaluation of self-exclusion programs including self-excluded patrons self-report indicates that it is common for breaches to occur and to go undetected. There are few systematic procedures in place to counter this.
A system program that is not capable of enforcing self-exclusion runs counter to the expectations of self-excluded patrons, counsellors, the media and community as well as venue staff, gaming operators and regulators. A failure to detect self-excluders who breach their agreement seriously undermines the program and may reduce the number of problem gamblers who register and the effectiveness for those who do register. Self-exclusion programs are limited in the extent to which they fail to cover gambling available at venues not included in the agreement. Studies have found that self-excluded individuals engage in gambling at venues they have not excluded from and in other forms of gambling to which bans do not apply. Some jurisdictions have introduced self-exclusion agreements that cover multiple gambling venues and multiple forms of gambling including EGMs, casino games and Internet gambling.

Recommended Elements of Self-exclusion
Based on the available evidence, the following elements are recommended to be included in all self-exclusion programs: • Clear information about self-exclusion programs and promotion of programs to increase utilisation. Promotion should include venue-based strategies as well as information provided in the general community and through health and mental health centres, legal offices, and other relevant support services. Relevant professional should all be informed about the program so that they may refer clients as appropriate. • To maximise the uptake of self-exclusion agreements, individuals should be able to enact self-exclusion agreements directly at gaming venues and also when off-site. The registration process should be conducted in a respectful manner, in a private setting and should be relatively quick and easy. If registration is perceived to be too difficult or complex, this may act as a barrier to self-exclusion. Self-exclusion agreements should be able to be enacted with treatment providers, counsellors or other suitable health and legal professionals. Self-exclusion agreements may also be initiated by individuals using Internet-based protocols that include webcams for photograph identification, by visiting suitable government agencies or mailing in applications and photographs.
• Programs must offer a range of time periods for exclusion. There is a lack of empirical evidence to suggest what length of ban is most effective in assisting individuals to control their gambling. Although longer bans may be more effective in providing individuals with the time needed to overcome their gambling-related problems, these may deter some individuals from registering for programs. It is suggested that selfexclusion agreements be a minimum of 6 months to allow individuals sufficient time to enter treatment if desired or deal with their gambling problems. • Operators must remove all self-excluded persons from the mailing list and stop offering incentives during the period of self-exclusion. Individuals must not be reinstated on any mailing lists even after their self-exclusion period has expired unless they have specifically applied and been approved to re-enter the venue. • Resources must be provided to assist self-excluded persons take further steps to control their gambling including appropriate information and education resources as well as referrals to formal counselling, treatment or self-help programs. Not all people seeking self-exclusion want or are receptive to professional help (Ladouceur et al. 2000Tremblay et al. 2008). Hence, mandatory counselling may not be effective and may deter some individuals from entering into self-exclusion agreements. Internet-based treatment and self-help interventions may be particularly useful for those who would benefit from formal help, but are unwilling or unable to attend in-person sessions (Gainsbury and Blaszczynski 2011a, b). • Operators must take active steps to identify and remove self-excluded persons who return to gaming facilities. It is reasonable to require that all patrons must show appropriate identification before they are admitted to gaming venues to prevent breaches. This would enable computerised identification checks for enforcement of self-exclusion. Checking identification has become part of standard practices for many businesses (e.g., buying alcohol, boarding planes), is accounted for in customer and queue management and is accepted by consumers. • A reinstatement process should occur before the self-excluded individual is permitted re-entry into gaming facilities. When the finishing date of their ban is approaching, self-excluded individuals should be contacted with appropriate information and details should be provided of what is required to reinstate entry. A re-entry session may be held and run by an external organisation to avoid self-excluders having to visit gaming venues. Alternatively, individuals should be able to extend their ban for a renewed time period and have a new photograph taken to ensure that it is current. Again, this process should be provided away from gaming venues to avoid triggering gambling behaviour and referrals should be provided for treatment services. • Operators must have an effective training program for all staff who have a role in enforcing the self-exclusion program, including refresher training. The training program should include how to identify individuals who may have a gambling problem and how to adhere to all components of the self-exclusion program. • Self-exclusion programs should be monitored and evaluated to ensure programs are effective, have no unintended harmful consequences and are conducted in compliance with the required processes. The limitations of the current literature and few evaluative studies conducted highlight the need to develop a systematic process for the evaluation of self-exclusion programs. A transparent system of monitoring and auditing a selfexclusion program would reduce any perceived conflict of interest by operators and clearly demonstrating a programs' effectiveness may increase utilisation.

Conclusions
Self-exclusion programs are an essential part of any harm-minimisation strategy offered by a gaming operator or jurisdictional regulator. Although self-exclusion programs have been in use since 1996, there are remarkably few comprehensive evaluative studies that have investigated the impact of these programs and the elements that should be included to maximise benefits. Despite the severe limitations to the available literature, there is some evidence that self-exclusion programs generally provide benefits to problem gamblers in terms of reduced gambling behaviour and reduction of problem gambling severity. There is also evidence of improved psychological functioning and perceptions of control over gambling behaviour. However, existing self-exclusion programs are under-utilised and do not appear to be effective in preventing gamblers from breaching agreements or gambling on non-restricted activities.
Regulation mandating operators to offer a self-exclusion program to patrons is expected to enhance the strength of this program by increasing the power to penalise individuals and operators who do not comply with the standards mandated or breach agreements. This may increase the effectiveness of programs as well as the perception of such strategies, which may in turn increase utilisation rates. Although there is no 'gold standard' program that can be implemented, it is important for any regulator to consider the existing evidence and make every attempt to implement minimum standards that are sufficient to ensure a programs feasibility. Self-exclusion programs should be flexible to accommodate the needs of individual gamblers, but must be recognised as a severe form of pre-commitment intended for those who are unable to control their own gambling behaviour. As such it must be sufficiently powerful to uphold self-exclusion agreements to the highest standard that can be reasonably expected in order to offer the maximum benefits and protection for individual self-excluded gamblers. It is expected that such programs would require constant evaluation, monitoring and modification as necessary and in line with developing technological capabilities. Further research is needed to develop more effective self-exclusion programs and also to consider tailored programs that are more relevant to individuals, for example, based on gender, age and cultural differences. Any jurisdiction that allows the provision of gambling should consider self-exclusion programs carefully in order to provide a duty of care to gamblers.