Dr. Robert Williams Faculty of Health Sciences & Alberta Gambling Research Institute University of Lethbridge, Alberta, Canada November 2015

Transcription

1 ETIOLOGY OF PROBLEM GAMBLING: Results from the Major Canadian Longitudinal Studies Dr. Robert Williams Faculty of Health Sciences & Alberta Gambling Research Institute University of Lethbridge, Alberta, Canada November 2015

2 Previous Longitudinal Studies of Gambling and Problem Gambling ~30 studies Some just looking at gambling (G) some just looking at problem gambling (PG) Some just looking at covariation of G or PG with other variables over time (rather than determining which variables actually predict subsequent PG)

3 Predictors of Subsequent Gambling from Prior Longitudinal Studies Substance use or misuse (6 studies) Impulsivity (4 studies) Male (2 studies) Financial problems (2 studies) 1 Study: depression, younger age, school problems, delinquency, moral disengagement, lower parental monitoring, greater social support

4 Predictors of Subsequent Problem Gambling from Prior Longitudinal Studies Alcohol problems (6 studies) Depression (6 studies) Impulsivity (5 studies) Tobacco use (5 studies) Less education and/or poor school performance (4 studies) Antisociality and/or conduct disorder (4 studies) Prior problem gambling and/or subclinical problem gambling (3 studies) Male (3 studies) Drug use or abuse (3 studies) Significant life events (3 studies)

5 Predictors of Subsequent Problem Gambling from Longitudinal Studies Stress and/or emotional distress (3 studies) Financial or employment concerns (3 studies) Poor health (3 studies) Frequency of gambling (2 studies) Use of Electronic Gambling Machines (2 studies) Anxiety (2 studies) Early onset of gambling (2 studies) Immigrant (2 studies) 1 study: internet gambling, parental gambling, risk-taking, race track gambling, family history of antisociality, antisocial peers, rebelliousness, alienation, one parent families, family conflict, greater social networking, aggression, novelty seeking, obesity, thoughts about winning

6 Limitations of these Longitudinal Studies Often very circumscribed demographic (e.g., just youth, just elderly, just casino employees). Often very small sample sizes and/or small number of people who become problem gamblers during course of the study. Often very short time span and/or a small number of assessments. Usually short questionnaire examining only small subset of variables potentially involved in etiology of problem gambling. Usually poor retention rates with differentially higher attrition for certain demographic groups.

7 Leisure, Lifestyle, Lifecycle Project (LLLP) Funded by Alberta Gambling Research Institute ($2.3 million) Alberta adults from 4 regions of Alberta approximating the Alberta population 29% oversampled for at risk characteristics 4 comprehensive assessments months apart Very similar questionnaire to QLS 2 3 hours per assessment Telephone interview (Assessment 1) + self-administered (online &/or paper & pencil) Dependent variable: score of 5 or higher on Canadian Problem Gambling Index (CPGI) (Ferris & Wynne, 2001) 76.2% retention rate

8 Quinte Longitudinal Study (QLS) Funded by Ontario Problem Gambling Research Centre ($3.1 million) Ontario adults from Quinte Region in southeastern Ontario, Canada 26% oversampled for at risk characteristics 5 comprehensive annual assessments Demographics, gambling, physical health, mental health, substance use, stressors, personal values, social functioning, personality, leisure activity, intelligence (135 variables) hrs per assessment self-administered online or via paper & pencil Dependent variable: problem or pathological gambler on Problem and Pathological Gambling Measure (PPGM) (Williams & Volberg, 2014) 93.9% retention rate

9 QLS & LLLP Research Questions 1. What is the natural stability of gambling and problem gambling over time? 2. What variables best predict future problem gambling? 3. What etiological model of problem gambling derives from these findings? 4. What are the implications of these results for the prevention of problem gambling?

10 Stability of Non-Gambling over Time (QLS) N = 280; each row represents an individual About 50% continued to be Non-Gamblers through 5 years: moderately stable category Not uncommon to transition to Recreational Gambling, but most revert back to Non-Gambling Very uncommon to transition to At Risk or Problem Gambling (only 1% became Problem Gamblers) Non-Gambler Recreational Gambler At Risk Gambler Problem Gambler

11 Stability of Recreational Gambling over Time (QLS) N = 2786; each row represents 25 individuals 70% continued to be Recreational Gamblers throughout the 5 years: stable category 13% transitioned into Non- Gambling and 10% into At Risk Gambling 5% became Problem Gamblers at some point Non-Gambler Recreational Gambler At Risk Gambler Problem Gambler

12 Stability of At Risk Gambling over Time (QLS) N = 481; each row represents an individual 1 Year the modal duration. Only 7% continued to be At Risk Gamblers throughout the 5 years. Most transitioned back into Recreational Gambling: unstable category 15% became Problem Gamblers at some point Non-Gambler Recreational Gambler At Risk Gambler Problem Gambler

13 Stability of Problem Gambling over Time (QLS) N = 236; each row represents an individual 1 year the modal duration, occurring in about 51% of people. Only 19% continued to be Problem Gamblers throughout the 5 years: unstable category Recovery rates high, but relapse rates also high (40% relapse within 3 years after recovery) Non Problem Gambler Problem Gambler

14 Stability of Pathological Gambling over Time (QLS) N = 88; each row represents an individual 64% have durations between 2 5 years, but only 30% Disordered Gamblers throughout the 5 years: moderately unstable category Recovery rates lower than problem gambling and relapse rates higher than problem gambling Non Problem Gambler Problem Gambler Pathological Gambler

15 Univariate Prediction of Future Problem Gambling Examined strength and consistency of variables as predictors of future problem gambling across: Time Periods 1 time period: A1 A2; A2 A3; A3 A4; A4 A5 2 time periods: A1 A3; A2 A4; A3 A5 3 time periods: A1 A4; A2 A5 4 time periods: A1 A5 and Data Sets QLS and LLLP

16 Univariate Prediction of Future Problem Gambling No single variable overwhelmingly present in future problem gamblers and absent in future non-problem gamblers. Rather, many different variables involved, each increasing risk to some extent and present to differing degrees in future problem gamblers. However, certain categories of variables more predictive and stronger variables within categories.

17 Univariate Prediction of Future Problem Gambling Gambling-Related Variables is category most robustly predictive of future problem gambling Being At Risk or Problem Gambler single best predictor of future problem gambling Intensity of gambling involvement 2 nd best predictor (i.e., total gambling expenditure, overall frequency, total time spent, number of formats played) Higher frequency of involvement in continuous forms (i.e., EGMs, casino table games, instant lotteries) 3 rd best predictor Other strong predictors: big win in past year; gambling a top leisure pursuit; family or friends regular or problem gamblers; gambling to escape or to win money ; more gambling fallacies; Internet gambling; proximity to EGM venues

18 Univariate Prediction of Future Problem Gambling Personality next most important category predictive of future problem gambling Impulsivity strongest personality predictor, and one of the strongest predictors across all categories Other fairly strong personality predictors: Vulnerability (to stress) Lower agreeableness Lower conscientiousness

19 Univariate Prediction of Future Problem Gambling Mental Health next most important category predictive of future problem gambling Depression strongest predictor in this category Other fairly strong mental health predictors: Anxiety-related disorders Substance abuse Having a behavioural addiction Lifetime history of mental health problems or addiction to drugs/alcohol

20 Other fairly strong and/or consistent predictors: More stressful events in past year Lower IQ Lower educational attainment Lower happiness Higher stress History of child abuse Antisocial traits Univariate Prediction of Future Problem Gambling Physical disability and/or poorer physical health

21 Multivariate Prediction of Future Problem Gambling Many univariate predictors not significant in multivariate prediction due to overlapping predictive power Almost all multivariate predictors were gambling-related: Being At Risk or Problem Gambler (strongest predictor) Big win in past year Frequency of EGM and/or casino table game participation Family members being regular gamblers Close friends/family with gambling problems Gambling to escape or win money More gambling fallacies Gambling a top leisure pursuit Engaging in larger number of gambling formats

22 Multivariate Prediction of Future Problem Gambling Only non-gambling related variables robustly adding multivariate predictive power were: Impulsivity Having a behavioural addiction (e.g., shopping, sex, video games, exercise) Lifetime history of addiction to drugs or alcohol Family history of mental health problems

23 Variables Predictive of First Onset Problem Gambling vs Continuation vs Relapse Almost all gambling-related variables predictive of first onset problem gambling. Exceptions being proximity to EGM venues and being At Risk or Problem Gambler, which were more predictive of continuation and relapse. Several personality, mental health, stress-related, cognitive, and physical health variables also implicated in first onset problem gambling. However, in general, personality, mental health, stress-related, cognitive, and physical health variables more strongly implicated in problem gambling continuation and relapse.

24 Proximal vs Distal Predictors Most predictors create risk for all future time periods, rather than some creating imminent risk and others creating distant risk. However, a few variables almost always precede problem gambling and are stronger predictors of imminent problem gambling than others: Intensive gambling involvement the strongest and most consistent predictor Other strong and consistent predictors of imminent problem gambling: Having big win in past year Gambling being a favourite leisure pursuit Impulsivity Depression

25 Proximal vs Distal Predictors No variables only predicted problem gambling in distant future. However, all significant predictors invariant over time create long-term risk: Personality (impulsivity, vulnerability, lower agreeableness, lower conscientiousness) Family history of heavy gambling and/or problem gambling Lower educational attainment Lower intelligence Lifetime history of mental health problems or drug/alcohol addiction Child abuse

26 Subjective Belief vs Objective Predictors (QLS) All problem gamblers asked What do you think caused your gambling problems? Only limited overlap between these open-ended reports and objective predictors. Most problem gamblers identified singular cause, whereas empirical results indicate many variables. Self-reported causes focused on psychological, motivational, and social influences (e.g., gambling to escape or to win money, boredom, stress/depression, social pressure to gamble). Although self-reported causes validated by empirical results, problem gamblers less aware of broader contextual determinants: past history of gambling problems, family history of gambling, engagement in continuous forms, big wins, gambling fallacies, personality, substance abuse, mental health problems.

27 Etiological Model Biopsychosocial etiology Multiple risk and protective factors Particular pattern of risk and protective factors different between problem gamblers, but many of the strongest risk factors tend to be fairly prevalent

28 Etiological Model: Heavy Gambling Final Common Pathway High levels of gambling expenditure, frequency, time, number of formats, and/or involvement in continuous forms creates greatest direct risk for problem gambling, as it immediately precedes problem gambling in large majority of cases. Heavy gambling also increases likelihood of big win, which is an important independent risk factor for problem gambling.

29 Etiological Model: Recovery & Relapse Common Recovery from problem gambling common, as modal problem gambling episode duration is only one year. Relapse back to problem gambling also common, with past history of problem gambling being strongest predictor of relapse and problem gambling continuation.

30 Etiological Model

31 Implications for Prevention of Problem Gambling 1. No silver bullet to prevent problem gambling. Rather, wide array of initiatives needed to address the multifaceted biopsychosocial etiology. Effectiveness of any single initiative will be modest, but coordinated efforts can have synergistic effects. 2. Generic school-based prevention programs targeting wide range of problems are needed.

32 Implications for Prevention of Problem Gambling 3. Effective treatment of substance abuse and mood disorders will reduce problem gambling incidence. Because of even more important role in continuation and relapse, comorbid mood disorders and substance abuse need to be routinely evaluated and treated in all problem gamblers. 4. Many risk factors have biological basis, making them difficult to address. However, people with these biological vulnerabilities often concentrated in poorer neighborhoods. Thus, not placing gambling opportunities in these neighborhoods one way of addressing these vulnerabilities.

33 Implications for Prevention of Problem Gambling 5. Most of the modifiable risk factors are gambling-related: Continuous forms should be eliminated, reduced in number or constrained in how they operate. Gambling reward programs should be eliminated or used to reward responsible gambling. ATMs and smoking should not be allowed in gambling venues. Player pre-commitment should be mandatory. Automated intervention needed to alert players to risky behaviour. Reducing availability of gambling will have modest impact on incidence, and more important role in reducing relapse.

34 Implications for Prevention of Problem Gambling 6. Education needed to promote knowledge, motivation, and attitudes conducive to responsible gambling.: Gambling fallacies and inappropriate gambling motivations (to escape or to win money) particularly important. Signs of problem gambling, elevated risks of continuous forms, facilitative effect of associating with heavy gamblers and problem gamblers, normative levels of time and money on gambling, true odds of gambling games, gambling practices that increase risk, where to go for help, low risk guidelines for problem-free gambling. Predictive power of existing low-risk guidelines (that focus on gambling expenditure) significantly improved with addition of: past history of PG; higher frequency of EGM and/or casino table games; family members and/or close friends regular or problem gamblers; big win in the past year; high level of gambling fallacies; gambling to escape ; history of impulsivity.

35 For More Information Williams, R.J., Hann, R., Schopflocher, D., West, B., McLaughlin, P., White, N., King, K., & Flexhaug, T. (2015). Quinte Longitudinal Study of Gambling and Problem Gambling. Report prepared for the Ontario Problem Gambling Research Centre. February 20, el-guebaly, N., Casey, D.M., Currie, S., Hodgins, D.C., Schopflocher, D., Smith, G.J., & Williams, R.J. (2015). The Leisure, Lifestyle, & Lifecycle Project (LLLP): A Longitudinal Study of Gambling in Alberta. Final Report for the Alberta Gambling Research Institute. February 2015.