Scottish Health Survey 2017 - volume one: main report

9. Gambling Behaviour

Malin Karikoski

Summary

• In 2017 adults took part in an average of 1.5 gambling activities in the past year.
• Excluding the National Lottery completely, 12% of all adults had participated in online gambling, with men more likely than women to do so (18% of men, 6% of women).
• Men took part in a wider range of activities than women; of the 19 different gambling activities 7 were undertaken by more than 10% of men whereas only 3 different gambling activities were undertaken by over 10% of women.
• Overall, more than 1 in 10 (12%) adults participated in four or more gambling activities in the last year with men more likely to do so (17% of men compared with 7% of women).
• The youngest and oldest age groups had the lowest gambling participation rates (51% of those aged 16-24 and 44% of those aged 75 and over) with participation highest among those aged 45-54 (73%). This pattern was largely influenced by national lottery only gambling – once this is excluded gambling activity was at its highest for those aged 25-34 (58%), gradually decreasing to 25% among those aged 75 and over.
• Adult gambling activity participation decreased from 70% in 2012 to 63% in 2017; largely driven by a decrease in National Lottery participation from 58% in 2012 to 46% in 2017.

• Prevalence of problem gambling was lowest among men aged 65-74 (0.2%). For women aged 16-24 and 45 and over there were no cases of problem gamblers among survey participants.
• Adults with a GHQ-12 score of 4 or more (indicative of a possible psychiatric disorder) were more likely to be classed as a problem gambler according to the DSM-IV scale than those with a GHQ- 12 score of 0 (1.7% for those with a score of 4 or more compared to 0.3 for those with a score of 0). This association was particularly evident for men – 3.4% of men with a GHQ-12 score of 4 or more were identified as problem gamblers compared with 0.6% for those with a score of 0.

9.1 Introduction

Gambling behaviour is increasingly a subject of public health and policy interest in Britain. In the past decade, the gambling landscape in Britain has changed significantly. The public now has access to an unprecedented variety of gambling apps, websites, online games and lotteries and online gambling behaviour is constantly evolving, such as following gambling companies on social media^[1].

In addition, the volume of gambling advertisement to which the public is exposed has increased drastically in recent years. A 2013 Ofcom report found that television advertisements for gambling had increased by 600% from 234,000 per year in 2007 to 1.39 million in 2012, with adults viewing approximately 630 adverts and under-16s exposed to an average of 211 adverts each^[2].

Gambling constitutes a serious public health concern. Problem gambling, which is defined as gambling behaviour that causes harm to the gambler and to those around them^[3] can contribute to a range of adverse mental and physical health issues, including experience of depression, insomnia, stress-related disorders as well as experience of comorbid disorders such as alcohol abuse or dependence^[4]. Annual statistics from GamCare^[5] identify the most commonly reported negative impacts of problem gambling as financial difficulties, anxiety and stress, and family and relationship difficulties.

The most severe form of problem gambling, pathological gambling, has been categorised as an impulse control disorder within the American Psychiatric Association's Diagnostic and Statistical Manual of Mental Disorders IV (DSM-IV) and has been included in the manual since 1980^[6].

9.1.1 Policy background

In Britain, gambling is positioned as a legitimate recreational and leisure activity with policy responsibility held by the British Department for Culture, Media and Sport. While gambling policy is a reserved matter, the Scotland Act 2016 gave Scottish Ministers limited powers to legislate on the number of Fixed Odds Betting Terminals in new betting shops only, but with the reduction of the maximum stake to £2, those very limited powers will effectively become null and void.

There is widespread recognition among policy makers, industry and health care professionals that, like alcohol consumption, some people who engage in gambling activity can experience harm. Unlike alcohol consumption, there are no specific policy targets relating to harm minimisation. The Gambling Act 2005^[7], which came into force in 2007, overhauled the way commercial gambling is licensed, advertised and regulated in the UK.

The Gambling Act 2005 contains three core licensing objectives. These are to:

• prevent gambling from being a source of crime or disorder, being associated with crime or disorder or being used to support crime,
• ensure that gambling is conducted in a fair and open way, and
• protect children and other vulnerable persons from being harmed or exploited by gambling^[8].

The final objective highlights the potential for some people who participate in gambling to experience harm as a result of their behaviour and states that these groups specifically should be protected.

In 2007, the British Medical Association highlighted the insufficient treatment facilities available for gambling problems and argued that services for problem gambling should be provided through the NHS, similar to those for drug and alcohol problems^[9]. The only structured NHS provision for gambling problems in the UK is the NHS National Problem Gambling Clinic in London. In recognition of this gap, the Government launched a consultation in 2017 including a proposal in support of Gamble Aware's ambition to open more clinics regionally and plans for a Northern NHS Gambling Clinic^[10].

In 2014, the Minister responsible for gambling, horse racing and the National Lottery announced a review of gambling advertising control, with youth and vulnerable groups a critical concern^[11]. Two years later (2016) the Government announced a call for evidence into gaming machines and social responsibility. Evidence on gambling advertising was also included but its focus was on code of conduct^[12]. The Committee of Advertising Practice remains open to amending the regulations around gambling advertising should new research provide a sufficient evidence base for this^[13].

In 2018, the Minister for Sport and Civil Society announced plans to reduce the maximum stake which can be placed on Fixed Odds Betting Terminals (FOBTs) from £100 to £2 to reduce the risk of gambling-related harm^[14].

9.1.2 Reporting on gambling in the Scottish Health Survey (SHeS)

This chapter presents estimates of past year participation in all forms of gambling in Scotland followed by estimates of problem and at-risk gambling according to two different measurement instruments, the DSM-IV and the Problem Gambling Severity Index (PGSI)^[15]. The relationship between problem and at-risk gambling and deprivation, mental health (using the General Health Questionnaire 12 – GHQ-12) and alcohol consumption are also examined.

The area deprivation data are presented in Scottish Index of Multiple Deprivation (SIMD) quintiles. Readers should refer to the Glossary at the end of this Volume for a detailed description of SIMD.

Supplementary tables on gambling are also published on the Scottish Health Survey website^[16]

9.2 Methods and Definitions

9.2.1 Gambling participation in the last year - definition and methods

All adult participants (aged 16 and over) were asked to report whether they had spent any money on nineteen different forms of gambling activity in the past 12 months. The activities presented ranged from buying tickets for the National Lottery draw to online betting and gaming. The range of activities presented reflected all forms of commercial gambling currently available in Scotland and also included betting or gambling privately with family or friends to capture informal gambling activity. In this chapter, gambling participation is defined as having participated in any one of these activities in the past 12 months. This definition also includes the requirement that the participant spent his/her own money on the activity. This was to ensure that those occasions where someone else placed bets or purchased lottery tickets with a participant's money were included.

The list of gambling activities and descriptions presented to participants reflected those used in the BGPS 2007 as closely as possible^[17]. Exceptions included the addition of 'playing poker in pub or club' and of 'betting on sports activities' (like football) to reflect the growing popularity of these activities since the 2007 study.

As with the BGPS series, questions were asked using a confidential self-completion format. This was to encourage more honest reporting of a (potentially) sensitive activity and to ensure maximum comparability with the BGPS. Everyone who had gambled at least once in the last year was also asked to complete two screening instruments to identify problem or risky gambling behaviour (see Section 9.2.2).

9.2.2 Problem gambling definition and measurement

Problem gambling is commonly accepted to involve 'gambling to a degree that compromises, disrupts or damages family, personal or recreational pursuits^[18]. Despite this, there is no definitive definition of problem gambling and many different instruments or 'screens' exist to identify and measure problem gambling (with over 20 different types in existence)^[19]. As yet, there is no agreed 'gold standard' instrument recommended for use in population surveys.

For this reason, it has been common practice in Great Britain to include two different screening instruments in population-based surveys of gambling behaviour. As the instruments tend to capture different types of people, using both together better reflects the broader range of issues associated with problematic gambling. The first of these is based on the fourth edition of the Diagnostic and Statistical Manual of Mental Disorders of the American Psychiatric Association (DSM-IV) and the second, the Problem Gambling Severity Index (PGSI)^[15], was developed in Canada specifically for use in population based studies. Both instruments have been widely used internationally and were the instruments of choice for the 2007 and 2010 BGPS. Since 2012, SHeS has included both the DSM-IV and the PGSI.

DSM-IV

The DSM-IV screening instrument contains ten diagnostic criteria ranging from chasing losses to committing a crime to fund gambling. Each DSM-IV item is assessed on a four-point scale, ranging from 'never' to 'very often'^[20]. Responses to each item are then dichotomised to show whether a person meets the criteria or not. A total score between zero and ten is possible. A threshold of meeting at least three of the DSM-IV criteria is used to define problem gambling. This cut-off point has been found to give good discrimination between criterion groups and has provided the closest match to prevalence estimated by alternative screens^[21]. Clinicians currently use an additional threshold of a DSM-IV score of five or more to represent pathological gambling^[15]. For a variety of reasons, this threshold is not presented in this chapter. Firstly, the number of people falling into this category would be too small to allow any detailed analysis to be carried out. Secondly, the term 'problem gambling' is preferred as it has less negative and medicalised conceptual issues associated with it than the term 'pathological gambling^xv.' Finally, it is likely that the label 'pathological gambling' will become obsolete as it has been renamed 'gambling disorder' in the recent publication of the DSM-V^[22]. The threshold and scoring criteria used to identify problem gamblers here are the same as those used in the BGPS series.

PGSI

The PGSI was developed for use among the general population rather than within a clinical context and was tested and validated within a general population survey. The instrument consists of nine items ranging from chasing losses to gambling causing health problems and feeling guilty about gambling. Each item is assessed on a four-point scale: never, sometimes, most of the time, almost always. Responses to each item are given the following scores: never = zero; sometimes = one; most of the time = two; almost always = three. Scores for each item are summed to give a total score ranging from zero to 27. A score of eight or over on the PGSI represent problem gambling. This is the threshold recommended by the developers of the PGSI and the threshold used in this report. The PGSI was also developed to give further information on sub-threshold problem gamblers. PGSI scores between three and seven are indicative of 'moderate risk' gambling^[23] and scores of one or two are indicative of 'low risk' gambling. As with the DSM-IV, the PGSI thresholds and scoring mechanisms used in SHeS are the same as those used in the BGPS.

Creating problem gambling scores

To produce problem gambling prevalence rates among all adults aged 16 and over, all non-gamblers were allocated a score of zero in both the DSM-IV and the PGSI screens. To be included in the final analysis for each instrument, participants were required to have answered at least five of the DSM-IV questions or at least four of the PGSI questions. Those who answered less than this were only included in the final analysis if their responses to the answered questions scored them as a problem gambler. Overall, around 10% of eligible adults did not have a valid DSM-IV or PGSI score for the 2016-2017 combined dataset and around 9% of eligible adults do not have a valid DSM-IV or PGSI score for the 2014-2017 combined dataset. This should be borne in mind when reviewing these results.

9.2.3 General Health Questionnaire 12 (GHQ 12)

GHQ-12^[24] is a standardised scale which measures mental distress and mental ill-health. There are 12 questions which cover concentration abilities, sleeping patterns, self-esteem, stress, despair, depression, and confidence in the past few weeks. For each of the 12 questions one point is given if the participant responded 'more than usual' or 'much more than usual'.

Scores are then totalled to create an overall score of zero to twelve. A score of four or more (described as a high GHQ-12 score) is indicative of a potential psychiatric disorder. Conversely a score of zero is indicative of psychological wellbeing. As GHQ-12 measures only recent changes to someone's typical functioning it cannot be used to detect chronic conditions.

9.3 Gambling Participation in the Last Year

9.3.1 Participation in gambling activities in last year, 2017, by age and sex

In 2017, six in ten (63%) adults (aged 16 and over) had gambled in the last 12 months, and in line with previous data^[25], men were more likely to have gambled than women (66% and 60% respectively).

Among all adults the most popular gambling activity was purchasing tickets for the National Lottery draw (46%). This was followed by buying scratchcards (22%), other lotteries (17%), betting on horse races (not online) (11%) and online betting with a bookmaker (10%). The prevalence of each of the other gambling activities asked about in the survey was 7% or less.

A total of 45% of all adults had participated in gambling activities excluding National Lottery only play (50% of men, 41% of women). Excluding the National Lottery, 12% of all adults had participated in online gambling (18% of men, 6% of women).

Four in ten (43%) women bought tickets for the National Lottery draw in the last year, making it the most popular gambling activity among women in 2017. This was followed by purchasing scratchcards (20%), other lotteries (18%), bingo (not online) (9%) and horse races (not online) (8%). The prevalence of each of the other gambling activities was 4% or less.

Among men, after the National Lottery (48%) and scratchcards (23%), online betting with a bookmaker were the most popular gambling activities (17%), followed by other lotteries (15%), horse races (not online) (13%), sports events (not online) (13%) and slot machines (10%). The remaining gambling activities had a prevalence of between 7% (football pools) and 1% (spread-betting).

Men took part in a wider range of gambling activities than women. Of the 19 gambling activities asked about, seven different gambling activities were undertaken by at least 10% of men (ranging from 10% of men using slot machines to 48% participating in the National Lottery), whereas only three different activities had a prevalence of over 10% for each activity among women (43% participated in the National Lottery, 20% had bought scratchcards and 18% had played in other lotteries).

As evident in Figure 9A, men and women prefer different types of gambling. Men were more likely to gamble online (excluding National Lottery) than women (18% and 6% respectively). This pattern is clear for online betting with a bookmaker, which was the third most popular activity among men (17%) and sixth most popular among women (4%).

Gambling activity participation in the past 12 months varied by age, with the youngest and oldest age groups having the lowest gambling participation rates (51% of those aged 16-24 and 44% of those aged 75 and over), with participation being highest among those aged 45-54 (73%). Age-related patterns of gambling participation were similar for men and women as demonstrated in Figure 9B.

These patterns are highly influenced by National Lottery only gambling activity. By looking at any gambling activity excluding National Lottery only play, different participation patterns by age emerge. Gambling activity (excluding National Lottery only) was at its highest for those aged of 25-34 (at 58%) and then gradually decreased as age increased down to 25% among those aged 75 and over.

Different patterns by age were found for men and women for gambling (excluding National Lottery only). For men this peaked among those aged 25-44 (61-66%) whilst for women it was more evenly spread across the age groups, being highest among those aged 25-54 (45-51%) see Figure 9C. Figures 9A, 9B and 9C, Table 9.1

9.3.2 Number of gambling activities undertaken in last year, 2017, by age and sex

In 2017, adults took part in 1.5 gambling activities on average; this was higher for men (1.8) than for women (1.2). The mean number of gambling activities was higher among those in the younger age groups for both men and women (see Figure 9D). The number of different gambling activities participated in peaked among those aged 25-34 (2.5 for men and 1.5 for women). The mean number of gambling activities then declined, with those aged 75 and over having the lowest mean number of activities (0.9 for men and 0.6 for women).

Young men were far more likely to engage in greater amounts of gambling activities than young women (2.1-2.5 mean number of activities for men aged 16-44 compared with 1.2-1.5 for women in the same age group. The difference between men and women was narrower among older age groups.

Overall, 12% of adults participated in four or more gambling activities in the last year with men more likely to do so (17% of men compared to 7% of women). Figure 9D, Table 9.2

9.3.3 Number of gambling activities undertaken in last year, since 2012

Gambling activity participation for adults was lower in 2017 than in 2012 (63% compared with 70%) with levels ranging between 65% and 68% in the intervening years. This is consistent with participation in the National Lottery falling from 58% in 2012 to 51% in 2014, rising to 53% in 2015 before dropping again to 46% in 2017. Similar patterns were found for men and women. However, men have been consistently more likely than women to gamble since the start of data collection in 2012, with the gap ranging between 6-10 percentage points over the survey years, see Figure 9E.

The trend pattern was less clear for any adult gambling activity excluding National Lottery only with the highest levels of gambling participation in 2015 and 2016 (both 49%). The level in 2017 (45%) was the same as in 2012. Gambling participation rates excluding National Lottery fluctuated over the time period for both men and women.

Despite gambling activities overall, as well as National Lottery as an individual activity being at the lower end of the scale in 2017, online gambling participation has remained at 12% since 2015, having steadily risen from 7% in 2012. A similar pattern was found in both men and women. Figure 9E, Table 9.3

9.4 Problem Gambling

9.4.1 DSM-IV and PGSI scores in the last year, 2016-2017 combined, by age and sex

Based on DSM-IV scoring, 0.8% of all adults in 2016-2017 (combined) were identified as problem gamblers. Men were significantly more likely to be problem gamblers than women (1.4%and 0.2% of women respectively). Based on PGSI scoring 0.5% of all adults had a score which indicated that they were a problem gambler. Similarly to the DSM-IV scoring, the rates for PGSI were significantly higher for men (1.0%) than women (0.1%).

Prevalence of problem gambling also varied significantly by age. According to the DSM-IV scores problem gambling increased from 0.8% among those aged 16-24 to the highest rates of 1.6% and 1.5% among those aged 25-34 and 35-44 respectively before declining to 0.6% among those aged 45-54, 0.4% for those aged 55-64 and further still to 0.1% among those aged 65 and over. Due to the low prevalence of problem gambling among women, this pattern is largely driven by men (although problem gambling was evident for women in the same age groups as the highest prevalence was observed among men, ages 25-44).

From PGSI scoring, problem gambling was also significantly associated with age. Adults aged 16-54 were more likely to be problem gamblers than those aged 55 and over (0.6-0.8% compared with 0.0-0.4%). Similarly to the pattern identified for DSM-IV scoring, the pattern by age for PGSI scores was driven mainly by men.

Using PGSI scoring, in 2016 to 2017 combined, 2.5% of adults were low risk gamblers and 1.0% moderate risk gamblers. Men were more likely than women to be low or moderate risk gamblers (4.2% and 1.8% respectively for men compared with 0.9% and 0.3% respectively for women). Moderate and low risk gambling also varied significantly with age. Prevalence peaked for both moderate and low risk gambling for the 25-34 and 35-44 age groups (moderate risk: 2.3% and 2.2% respectively, low risk: 4.0% and 3.9% respectively) and declined gradually with age with the lowest prevalence observed among those aged 65 and over. Table 9.4

9.4.2 DSM-IV and PGSI scores in the last year, 2016-2017 combined, by area deprivation

Based on DSM-IV scores, there was a significant difference in problem gambling prevalence between those living in the most deprived areas (1.5%) and those living in the least deprived areas (0.5%). However the pattern across the deprivation quintiles was not clear with prevalence lowest among those in the second most deprived quintile (0.3%). This pattern was largely driven by men due to the lower prevalence of problem gambling among women.

Similarly PGSI scores which indicate problem gambling differed significantly between those living in the most deprived areas (1.0%) and the least deprived areas (0.2%). A similar pattern was observed for low risk and moderate risk gambling prevalence which was significantly higher among those living in the most deprived areas compared with the least (low risk: 4.5% compared with 2.0%, moderate risk: 1.6% compared with 0.6%). Table 9.5

9.4.3 DSM-IV and PGSI scores in the last year, 2014-2017 combined, by GHQ-12

Using DSM-IV scoring, 1.7% of adults with a GHQ-12 score of four or more were identified as problem gamblers, compared to 0.3% of those with a GHQ-12 score of zero. This pattern was largely driven by men (3.4 of men with a GHQ-12 score of four or more were problem gamblers, compared to 0.6% of those with a GHQ-12 score of zero).

Using PGSI scores, 1.2% of adults with a GHQ-12 score of four or more were identified as problem gamblers compared to 0.4% of those with a GHQ-12 score of zero. A similar pattern was found among those identified as moderate risk gamblers (2.0% of those with a GHQ-12 score of four or more compared to 0.7% of those with a GHQ-12 score of zero). There was no significant difference in prevalence of low risk gambling among those with a GHQ-12 score of four or more and 0. However there was a significant difference between those with a score of 0 and a score of 1-3 (2.2% and 3.3% respectively).

Similar patterns were found for both men and women. Table 9.6

9.4.4 DSM-IV and PGSI scores in the last year, 2014-2017 combined, by alcohol consumption

Problem gambling prevalence using the DSM-IV scale was marginally significantly associated with alcohol consumption with 0.3% of non-drinkers identified as problem gamblers compared with 1% of hazardous or harmful drinkers. However the differences in problem gambling prevalence using PGSI scores by alcohol consumption were not statistically significant.

Hazardous / harmful drinkers were more likely than non-drinkers to be identified as moderate risk gamblers by PGSI scores (2.4% compared to 0.8%) as well as low risk gamblers using PGSI scores (3.9% compared to 2.4%). Table 9.7