INTRODUCING THE CANADIAN PROBLEM GAMBLING INDEX. Harold J. Wynne, Ph.D. Wynne Resources Edmonton, Alberta Canada

Transcription

1 INTRODUCING THE CANADIAN PROBLEM GAMBLING INDEX Harold J. Wynne, Ph.D. Wynne Resources Edmonton, Alberta Canada January 2003

2 Page 2 INTRODUCING THE CANADIAN PROBLEM GAMBLING INDEX Harold J. Wynne, Ph.D. May 2002 BACKGROUND The development of the Canadian Problem Gambling Index (CPGI) was the result of a collaborative, three-year research venture ( ) managed by the Canadian Centre on Substance Abuse (CCSA) and funded by the Canadian provinces. The goal was to develop a new, more meaningful measure of problem gambling for use in general population surveys in Canada, one that reflected a more holistic view of gambling within a social and community context. The CPGI was developed as part of a larger research project entitled Measuring Problem Gambling in Canada. The Inter-Provincial Task Force on Problem Gambling was formed to guide the three-year study, and this group was comprised of government and agency representatives from across Canada (see Appendix 1 for a list of reps). The CCSA research team members included Dr. Harold Wynne, Jackie Ferris, Dr. Eric Single, and Dr. Nigel Turner and the following expert panel of Canadian and international gambling researchers advised the team at various stages of the project: Robert Ladouceur, Ph.D., Laval University, Quebec City Jeff Derevensky, Ph.D., McGill University, Montreal Rina Gupta, Ph.D., McGill University, Montreal Ron Frisch, Ph.D., University of Windsor, Ontario Sue Fisher, Ph.D., Cornwall, England Henry Lesieur, Ph.D., Pawtucket, Rhode Island Durand Jacobs, Ph.D., Loma Linda University, Redlands, California Randy Stinchfield, Ph.D., University of Minnesota, Minneapolis Ken Winters, Ph.D., University of Minnesota, Minneapolis Robert Wildman, Ph.D., Reno, Nevada Alex Blaszczynski, Ph.D. University of New South Wales, Sydney, Australia Michael Walker, Ph.D., University of New South Wales, Sydney, Australia The CPGI instrument resulted from (a) a review and synthesis of the most current gambling research available, and (b) expert opinion from internationally renowned gambling researchers, and it draws on the measures that have been used in the past for many of its key items. Consequently, the CPGI is less a new instrument and more an evolution of older measures, and it is based on the following operational definition of problem gambling developed by the research team: Problem gambling is gambling behaviour that creates negative consequences for the gambler, others in his or her social network, or for the community.

3 Page 3 METHODOLOGY The three-year Measuring Problem Gambling in Canada research project was conducted in two phases--a development phase and a testing phase. The development phase involved the clarification of the concept of problem gambling, the positing of an operational definition, and the designing of an instrument for measuring problem gambling in the general population. The second phase involved validity and reliability testing of the new instrument in a Canada-wide gambling prevalence study. The second phase of this project involved the fine-tuning, validity and reliability testing of the instrument developed in Phase I. This was accomplished by (a) conducting a pilot test of the face-valid instrument with 143 people; (b) following this pilot test with a general population survey of 3,120 Canadian adults; (c) re-testing 417 respondents from the general population survey, and (d) conducting clinical interviews with 148 respondents from the general population survey. The pilot test was conducted with three separate groups of about 50 persons each, intended to represent three potential sub-groups classified by the new instrument, namely: (1) a random, general population sample unlikely to have gambling problems (nonproblem gamblers); (2) a group of regular gamblers who may be at risk because of higher participation frequency and expenditure on gambling (at-risk gamblers); and (3) those who felt that they have a gambling problem at the severe end of the continuum (problem gamblers). This pilot test was used to determine if there were scored items from the facevalid instrument that did not discriminate between groups, and so could be eliminated in the next round of testing. Consequently, several items were deleted from the final survey instrument, but the South Oaks Gambling Screen (SOGS) and DSM-IV items were maintained in their entirety for comparative purposes. The final survey questionnaire, named the Canadian Problem Gambling Index, was administered to a random sample of 3,120 adults selected from the ten Canadian provinces. This large, general population survey was conducted for three main reasons. First, the large sample size allowed for fine-tuning the new instrument in terms of selecting items to be included/discarded in future iterations. Second, the survey provided a basis for comparison of the new instrument scored items with those from the DSM-IV and SOGS measures. Finally, the survey provided gambling/problem gambling prevalence information for the Nation and for individual Canadian provinces, and these data can be used as a benchmark for subsequent national or provincial studies. Although reliability can be assessed using a measure of internal consistency, and so could be calculated based solely on the general population survey, it was determined that a re-test of a small sub-sample from the general population survey would add to the rigor and credibility of the validation process. Consequently, a random sample of 417 respondents was re-called, and asked to complete the survey for a second time. Finally, to further test the validity of the new measure, clinical psychologists conducted telephone interviews with a sub-sample of 148 respondents from the general

4 Page 4 population survey. This was expected to provide confirmation of the classification scoring accuracy of the new instrument. CANADIAN PROBLEM GAMBLING INDEX The new instrument has been named the Canadian Problem Gambling Index (CPGI). The CPGI examines eighteen variables in four domains, and specific measurable indicators and corresponding questionnaire items are identified in Appendix 2. Domains of the CPGI Gambling involvement. The first CPGI section explores gambling involvement, with questions about (a) type of gambling activity, (b) the frequency of play, (c) duration of play, and (d) expenditure. The type of activity question is adjusted to suit the jurisdiction in which the CPGI is being used, and researchers are encouraged to identify as many gambling activities as are appropriate. These questions take very little time to administer, especially for those who participate in few gambling activities, which is the majority of respondents in a general population survey. Problem gambling assessment. The CPGI assesses two domains of problem gambling, namely (a) problem gambling behaviour, and (b) consequences of that behaviour for the individual or others. There are twelve items in these two domains, and nine of these are scored to determine problem gambling severity. This nine-item index is referred to as the Problem Gambling Severity Index (PGSI), and these items, along with the scoring algorithm, are included in Appendix 3. Depending on how a respondent scores on these nine PGSI items, he or she may be classified as being in one of four gambler sub-types, namely: (a) non-problem gambler, (b) low risk gambler, (c) moderate risk gambler, and (4) problem gambler (qualitative descriptions for each of these four sub-types are included in Appendix 4). Scoring the nine-item PGSI is crucial to the classification of gambler sub-types and, therefore, the PGSI items should not be removed or altered in any way. Correlates. The problem gambling correlates domain includes variables that further develop the profiles of gambler sub-types. Findings from the latest research in the field of gambling studies are reflected in these variables, and it is expected that new correlates will continue to be added by researchers. Reliability and Validity of the PGSI The psychometric properties of the Canadian Problem Gambling Index, including the nine-item Problem Gambling Severity Index, are described in detail in the Phase II report available on the Canadian Centre on Substance Abuse web site ( A brief summary of the reliability and validity of the PGSI follows.

5 Page 5 Reliability. Reliability is the ability of a measuring instrument to give consistent results on repeated trials. The PGSI was subjected to two types of reliability testing; namely, internal consistency and repeated measures methods. First, the Cronbach alpha reliability coefficient was computed, which provided an indication of the internal consistency of the PGSI measure. The alpha coefficient for the PGSI (0.84) was stronger than that for either the SOGS (0.81) or DSM-IV (0.76). The reliability of three sub-components of the PGSI was also examined (i.e., the two items retained from the SOGS, the two from the DSM-IV, and the five new PGSI items). The two SOGS items (0.71), two DSM-IV items (0.79), and five PGSI items (0.80) correlated relatively highly with their parent measures, respectively, although they did not perform reliably as two or five-item stand-alone measures. The second method for examining reliability was to repeat the survey by retesting the same people on the same three measures. To establish re-test reliability, 417 of the original 3,120 respondents were re-surveyed, and the Pearson Product-Moment correlation coefficient for the three measures included in the survey was calculated. In terms of re-test reliability, the DSM-IV (0.91) was strongest over time, followed by the PGSI (0.78) and SOGS (0.75). From these tests, it is apparent that the nine-item PGSI is a reliable measure of problem gambling. Validity. The validity of the PGSI was addressed in terms of content (face) validity, criterion-referenced validity, and construct validity. Content validity is the degree to which a test measures an intended content area. In this vein, the original PGSI went through several rounds of expert feedback, with twelve of the world s foremost gambling experts providing advice on the most appropriate content needed to discriminate problem gambling (i.e., domains, variables, measurable indicators and items). The expert consensus was that the five PGSI problem gambling behaviour and four adverse consequence items appeared to measure both the construct and operational definition of problem gambling very well. Subsequent statistical factor analysis of the PGSI confirmed that the nine items in the index load on one factor, which the researchers naturally labeled problem gambling. Consequently, it may be concluded that the PGSI has very good content validity as a measure of problem gambling. Criterion-referenced validity generally refers to validity that is found by correlating a measure with an external criterion of the entity being assessed. There are two types of criterion-referenced validity concurrent validity and predictive validity. Concurrent validity refers to validity assessed by correlating a measure with an alternative measure of the same phenomenon taken at the same point in time. Predictive validity refers to the ability of a measure to predict some attribute or characteristic in the future.

6 Page 6 In the Canadian national survey, the nine-item PGSI, the twenty-item SOGS, and the ten-item DSM-IV measures were administered to each of the 3,120 adult respondents. The PGSI was shown to have good concurrent validity because its scores correlated highly with those from both the SOGS (0.83) and the DSM-IV (0.83). In terms of predictive validity, the PGSI (0.48) also had a higher correlation with the clinical assessment interview results than either of the other two measures, although the correlation is only moderate. In view of these findings, it may be concluded that the PGSI has strong concurrent validity with the SOGS and DSM-IV, and moderate predictive validity with clinical interviews. Finally, the validation study examined PGSI construct validity, which refers to the degree to which the measurement agrees with the theoretical concept being investigated. In terms of theory, one might expect that those in the problem gambling group would spend more money on gambling, gamble more frequently, and be more likely to endorse the problem gambling correlates. When trend analysis was conducted on the data, differences were apparent by gambler sub-type on almost all of these indicators and correlates. Those in the problem gambling group do spend more, gamble more often, and endorse most of the correlates at higher rates than those in the at-risk or non-problem groups. In view of this, it may be concluded that the PGSI is a valid measure of the construct labeled problem gambling. Classification Accuracy In the context of screening, there are two other measures of validity that must be considered, namely sensitivity and specificity. Sensitivity refers to the ability of the test to identify correctly all screened individuals who actually have the condition. In contrast, specificity refers to the ability of the test to identify only non-diseased individuals who actually do not have the condition. A main challenge in validating the PGSI was to set cut-points along the continuum of scores (i.e., from 0 to 27 points) to maximize both the sensitivity and specificity of the measure. It was important to ensure that those classified as problem gamblers according to PGSI scores were true positives for this condition; moreover, it was also important that the cut-point ideally captured everyone who was a problem gambler (sensitivity). Similarly, cut-points needed to be set to ensure that individuals who were not problem gamblers were identified as not having this condition (specificity). Typically in setting cut-points, there is a trade-off between maximizing sensitivity versus specificity, and this was the challenge in validating the PGSI. In the final analysis, the cut-point set to discriminate the problem gambling sub-group (PGSI 8) had relatively good sensitivity, as it successfully (a) identified seven out of nine (78%) individuals who were classified in the clinical interviews as being problem gamblers, and (b) identified 83% of the individuals who scored on the DSM-IV diagnostic measure as being pathological gamblers. Specificity was 100% for all measures, and this means that the people identified by the DSM-IV and SOGS as having no gambling problem were also categorized as having no problem by the PGSI.

7 Page 7 In determining the cut-point that discriminates low-risk from moderate-risk gambler sub-types (i.e., PGSI = 2 vs. 3), the distribution of DSM-IV and SOGS scores, along with clinical assessment interviews that differentiated those deemed to be at low versus moderate risk, were taken into account. Issues of specificity and sensitivity have little meaning when applied to these at-risk sub-groups, as individuals therein do not have the problem gambling condition, per se; rather, they are deemed to be at greater/lesser risk for developing a gambling problem. CONCLUDING COMMENT A tenet of public health is that primary prevention of disease and socio-health disorders such as problem gambling is the best approach. However, if these health conditions cannot be prevented, then the next best strategy is early detection in asymptomatic, apparently healthy individuals. To facilitate early detection, screening measures are typically employed, and the nine-item PGSI was developed as such a measure to detect individuals in the general population who have a gambling problem, or who are at risk for developing a problem. It should be noted that screening differs from diagnosis, which is the process of confirming an actual case of a disease or health condition. As a result of diagnosis, a treatment intervention may be initiated. Diagnostic tests are used to follow-up positive screening test results, or to concurrently screen for and diagnose a health condition. In screening for and diagnosing problem gambling, the gold standard measures have been the South Oaks Gambling Screen and DSM-IV, respectively. Interestingly, the SOGS has also been used in research and clinical interventions to diagnose problem gambling, notwithstanding that it was developed as a screening instrument. Similarly, the DSM-IV has been used in population surveys to screen individuals for a gambling problem, despite its primary purpose being to diagnose the mental health disorder labeled as pathological gambling. The utility of both the SOGS and DSM-IV for screening and/or diagnosing problem gambling has been debated at length in the gambling literature. Given that severe problem gambling is a fairly robust health disorder, the SOGS, DSM-IV, and other related measures have been generally successful in screening for the most seriously disordered. However, these measures have had far less utility for classifying individuals who have sub-clinical symptoms, or who may be at some level of risk for developing a gambling problem. Like these other measures, the nine-item PGSI successfully classifies those problem gamblers who are most severely disordered. However, unlike other measures, the PGSI has greater classification accuracy for successfully identifying individuals who are at low or moderate risk for developing a gambling problem. Moreover, given that the PGSI is part of the larger CPGI measure, there is much more information gathered about the at-risk and problem gamblers experience. This wealth of CPGI gambling-related information results in a much more detailed picture of the relationship between gambling

8 Page 8 activities, problem gambling behaviour, adverse consequences, and other problem gambling correlates. Future CPGI-based research. Since the Canadian Problem Gambling Index was published in 2001, it has been integral to a growing body of research in the gambling studies field. This ever-increasing use of the CPGI bodes well for (a) the future refinement of the instrument, including the PGSI sub-index, and (b) the development of a growing statistical database that includes directly comparable research results. The following are known instances where the CPGI has been utilized in recent gambling research projects: Canadian CPGI-based research 1. Doiron, J. & Nicki, R. (1999). The Prevalence of Problem Gambling in Prince Edward Island. Prepared for the Prince Edward Island Department of Health and Social Service. Prince Edward Island. This is the CPGI baseline study of the prevalence of problem gambling in PEI. This research used an earlier, non-validated version of the CPGI (N=809). 2. Ferris, J. and Wynne. H. (February 2001). The Canadian Problem Gambling Index: Final Report. Report to the Canadian Inter-Provincial Advisory Committee. Ottawa, ON: Canadian Centre on Substance Abuse. This report contains information on the Canadian national survey that validated the newly developed CPGI (N=3,120). 3. Wiebe, J., Single, E, and Falkowski-Ham, A. (November 2001). Measuring Gambling and Problem Gambling in Ontario. Report to the Ontario Problem Gambling Research Centre. Submitted by the Canadian Centre on Substance Abuse and the Responsible Gambling Council (Ontario). This is the CPGI baseline study of the prevalence of gambling and problem gambling in Ontario (N=5,000). 4. Wynne, H. (January 2002). Gambling and Problem Gambling in Saskatchewan. Report to Saskatchewan Health. Submitted by the Canadian Centre on Substance Abuse. This is the CPGI baseline study of the prevalence of gambling and problem gambling in Saskatchewan (N=1,848) 5. Smith, G. and Wynne, H. (February 2002). Measuring Gambling and Problem Gambling in Alberta Using the Canadian Problem Gambling Index: Final Report. Report to the Alberta Gambling Research Institute.

9 Page 9 This is the CPGI baseline study of the prevalence of gambling and problem gambling in Alberta (N=1,804). 6. Patton, D., Brown, D., Dhaliwal, J., Pankratz, C., and Broszeit, B. (April 2002). Gambling Involvement and Problem Gambling in Manitoba. Report to the Addictions Foundation of Manitoba. This is an omnibus prevalence survey that examines gambling in various populations in Manitoba, including adolescent, adult, senior, women, and First Nation samples (n=3,119). The adult survey serves as a CPGI baseline study for this population as this instrument, as well as the SOGS, was utilized to discriminate gambler sub-types. 7. CPGI prevalence studies are presently being designed for the Canadian provinces of Quebec, British Columbia, and Nova Scotia. 8. Statistics Canada is presently conducting a Canadian national mental health study, and the survey instrument includes a module on problem gambling. A modified version of the CPGI, which includes the full PGSI, is the instrument that is being used in the survey s problem gambling module (N=35,000). 9. Wiebe and Single are presently examining the OPGRC problem gambling framework by plotting data from the Ontario CPGI study within the framework s four risk categories. 10. In an OPGRC-funded study of the separate and combined effects of cognitive behavioral therapy and medication on pathological gambling, Ravindran and Telner are measuring pre-post changes based on PGSI scores. 11. At the Royal Alexandra Hospital in Edmonton, Smith and Wynne are using the PGSI in an opportunistic screening protocol for health conditions with all nonurgent patients presenting in the emergency department. CPGI-based research in other countries 12. Gambling Policy Directorate (2002). Queensland Household Gambling Survey Report to the Queensland Government Treasury. This gambling/problem gambling prevalence telephone survey was conducted in September 2001 with a random sample of 13,082 Queensland, Australia residents over 18 years-of-age. The researchers chose to use both the CPGI and SOGS to identify gambler sub-types, and a comparative analysis concluded that, the results show that all problem gambling respondents who have been captured by the CPGI were also captured by the SOGS. Findings and conclusions were subsequently

10 Page 10 reported based on the CPGI classifications. The researchers concluded, The CPGI provides important information about gambling groups in the population. Research is warranted to identify the common characteristics of the Low Risk, Moderate Risk and Problem Gambling groups which might identify those at most risk of developing problem gambling. 13. Gambling Policy Directorate (2002). Problem Gambling Prevalence Survey Report to the Queensland Government Treasury. This report outlines the findings of research conducted into the incidence of problem gambling within a random sample of 178 male and female prisoners incarcerated in Queensland, Australia corrective services facilities. The CPGI was used to classify gambler subtypes because, according to the Gambling Policy Directorate, the Canadian Problem Gaming Index (CPGI) was utilised by Treasury in its Queensland Household Gambling Survey 2001, and In addition, the CPGI s definition of problem gambling, as referred to in Section 1 of this report is analogous to that which is currently utilised for policy purposes in Queensland. 14. Linnet is conducting a study in Denmark with pathological gamblers in treatment, and a translated version of the PGSI is being used to discriminate gambler subtypes. 15. McMillen is working with two Australian state governments that are contemplating using the PGSI to screen problem gamblers for treatment.

11 Page 11 APPENDIX 1 Inter-Provincial Task Force Members And Key Informants

12 Page 12 Name Title/Affiliation/Location Telephone Contact Via: New Brunswick: Jean Guy Le Blanc Gambling Counselor Carolyn Green Gambling Counselor FAX: Bob Jones Treatment Consultant/Coordinator Carol Harned Outpatient Counselor FAX: Dr. Alberto Barcelo Task Force Member Nova Scotia: Jerome McCoyne Addiction Services Ann McLean Field administrator Joel Baltzer Researcher, Gaming Control Commission Brian Wilbur Director, Drug Dependency Services Carrie Chambers Researcher, Sterling Research Tracey Shranz Researcher, Focal Research John Laroque Ontario: Carolyn Nutter ARF, VP Community Health and Education Nigel Turner ARF, psychologist, psychometrist Tony Toneatto ARF, psychologist, clinician Nina Littman-Sharpe Donwood, clinician, head of gambling treatment program Geoff Noonan Gambling counselor, Donwood Roger Horbay Gambling counselor, Donwood Jane Scott Gambling counselor, CFCG (Ontario) Dr. Stanley Debow Gambling counselor MAIL Colette Prevost Addiction Services of Nipissing PHONE Patrick Au Counselor, Chinese Family Life FAX: Services Lynn Eaton Gambling counselor, Grey Bruce FAX: Alcohol Assessment Wilf Bowering* Homewood Health Centre, head of gambling programs X291 FAX:

13 Page 13 Manitoba: Jamie Wiebe Researcher, AFM Ron Norton Professor, University of Winnipeg H.Uwinnipeg.ca Gerry Kolesar Supervisor, Gambling Programs AFM Saskatchewan: Bob Markosky Saskatchewan Health Steve Christensen Saskatchewan Health a Kevin Hanna Community Mental Health Services Jean Dunlop Saskatoon Health District, FAX: Addictions Don Osga Mental Health Jim Worrel Prince Albert Health District FAX: Maureen Boyko Addiction Services FAX: Alberta: Art Dyer David Hodgins Barry Andres* British Columbia: Jane Burke Kathie Neufeld Greg McQuarrie Thomas Chan Miki Hansen Manager, Service Monitoring & Research, AADAC Director of Research Addiction Centre, Foothills Hospital, Calgary Counselling Supervisor, AADAC Adult Services Consultant, women and problem gambling, Peardonville House Consultant, seniors and problem gambling, ADEPT Services Specialist, aboriginals and problem gambling, Northwest Alcohol and Drug Services Specialist, multi-cultural populations and problem gambling, Family Services of Greater Vancouver Director, Adult Addictions Services Branch, BC ca ca ab.ca X

14 Page 14 APPENDIX 2

15 Page 15 Canadian Problem Gambling Index For each of the items in the CPGI questionnaire, respondents are asked to respond in the past twelve (12) months. This past-year time frame does not apply to the following questions: 18, 19, 20, 21, 22, and 23. The response scales for each of the questionnaire items are as follows: Question 1 - Question 2 - Question 3 - Questions 4 & 5 - Questions 5 to 17 - Questions 18 & 19 - Questions 20 to 33 - yes; no daily; 2-6 times/week; about once/week; 2-3 times/month; about once/month; between 6-11 times/year; between 1-5 times/year; never in the past year record actual minutes and/or hours record actual dollar amount never; sometimes; most of the time; almost always strongly agree; agree; disagree; strongly disagree yes; no DOMAIN VARIABLES INDICATORS ITEMS AND QUESTION NUMBERS Type Gambling 1. Have you bet or spent money on (list of gambling activities)? activities Gambling Frequency Frequency of play 2. How often did you bet or spend money on (list activity: daily, weekly, monthly, yearly)? Involvement Duration Time at play/type/session 3. When spending money on (list activity), how many minutes/hours do you normally spend each time? Expenditure Money wagered monthly Largest amount wagered 4. How much money, not including winnings, did you spend on (list activity) in a typical month? 5. What is the largest amount of money you ever spent on (list activity) in any one day? Loss of Bet more than 6. How often have you bet more than you could really afford to lose? control could afford Bet or spent more than 7. How often have you bet or spent more money than you wanted to on gambling? wanted to Motivation Increase wagers 8. How often have you needed to gamble with larger amounts of money to get the same feeling of excitement? Problem Chasing Returning to win back losses 9. How often have you gone back another day to try to win back the money you lost? Gambling Behavior Borrowing Lying Borrow money or sold anything Lied to family members or others 10. How often have you borrowed money or sold anything to get money to gamble? 11. How often have you lied to family members or others to hide your gambling?

16 Page 16 DOMAIN VARIABLES INDICATORS ITEMS AND QUESTION NUMBERS Problem Felt problem 12. How often have you felt that you might have a problem with gambling? recognition Personal Consequences Wanted to stop, didn t think could Criticism 13. How often have you felt like you would like to stop betting money or gambling, but you didn t think you could? 14. How often have people criticized your betting or told you that you had a gambling problem, regardless of whether or not you thought it was true? Adverse Consequences Problem Gambling Correlates Social Consequences Feelings of guilt Negative health effects Financial problems 15. How often have you felt guilty about the way you gamble or what happens when you gamble? 16. How often has gambling caused you any health problems, including stress or anxiety? 17. How often has your gambling caused any financial problems for you or your household? Faulty Due for a win 18. After losing many times in a row, you are more likely to win. cognition after losses Having a 19. You could win more if you use a certain system or strategy. winning system First Remember a big 20. Do you remember a big WIN when you first started gambling? experiences win Remember big 21. Do you remember a big LOSS when you first started gambling? loss Family Family gambling 22. Has anyone in your family EVER had a gambling problem? problems problem Family alcohol 23. Has anyone in your family EVER had an alcohol or drug problem? or drug problem Co-Morbidity Gamble, drugs, 24. Have you used alcohol or drugs while gambling? alcohol Gamble when 25. Have you gambled while drunk or high? high Problem Felt alcohol/drug 26. Have you felt you might have an alcohol or drug problem? recognition problem Relieve pain Self-medication 27. If something painful happened in your life, did you have the urge to gamble? (gambling) Self-medication (alcohol) 28. If something painful happened in your life, did you have the urge to have a drink? Self-medication (drugs) 29. If something painful happened in your life did you have the urge to use drugs or medication? Stress Treated for stress 30. Have you been under a Dr s care because of physical or emotional problems brought on by stress? Depression Feelings of depression 31. Was there ever a time when you felt depressed for two weeks or more in a row? Suicide Suicide ideation 32. Have you ever seriously thought about committing suicide as a result of your gambling? Suicide attempts 33. Have you ever attempted suicide as a result of your gambling?

17 Page 17 APPENDIX 3 Problem Gambling Severity Index

18 Page 18 PROBLEM GAMBLING SEVERITY INDEX Scoring Algorithm and Questionnaire Items 1. PGSI Questionnaire Items Scored The 9 items (Q1-Q9) in the questionnaire below are scored. Score 1 for each response of sometimes, 2 for each most of the time, and 3 for each almost always. A score of between 0 and 27 points is possible. 2. Classification of Gambler Sub-Types There are four classification categories based on the following cut-points for PGSI scores: o 0 = non-problem gambler o 1-2 = low risk gambler o 3-7 = moderate risk gambler o 8+ = problem gambler The non-problem gambler group is separated into gamblers and non-gamblers as these sub-groups have quite different characteristics. 3. PGSI Scored Items by Category Dimension Variables Indicators PGSI Scored Items Problem Gambling Behaviour Adverse Consequences Loss of control Bet more than could afford 1. How often have you bet more than you could really afford to lose? Motivation Increase wagers 2. How often have you needed to gamble with larger amounts of money to get the same feeling of excitement? Chasing Return to win back losses 3. How often have you gone back another day to try to win back the money you lost? Borrowing Borrow money or sold anything 4. How often have you borrowed money or sold anything to get money to gamble? Problem 5. How often have you felt that you might have a problem with gambling? Felt problem recognition 6. How often have people criticized your betting or told you that you had Criticism a gambling problem, regardless of whether or not you thought it was Personal consequences Social consequences Feelings of guilt Negative health effects Financial problems true? 7. How often have you felt guilty about the way you gamble or what happens when you gamble? 8. How often has your gambling caused you any health problems, including stress or anxiety? 9. How often has your gambling caused any financial problems for you or your household?

19 Page 19 CANADIAN PROBLEM GAMBLING INDEX Problem Gambling Severity Index Questionnaire Items Q1. Thinking about the past 12 months, how often have you bet more than you could really afford to lose? Would you say: Never...1 Sometimes...2 Most of the time...3 Almost always...4 (DO NOT READ) Don't Know...8 (DO NOT READ) Refused / No Response...9 Q2. Thinking about the past 12 months, how often have you needed to gamble with larger amounts of money to get the same feeling of excitement? Never...1 Sometimes...2 Most of the time...3 Almost always...4 (DO NOT READ) Don't Know...8 (DO NOT READ) Refused / No Response...9 Q3. Thinking about the past 12 months, how often have you gone back another day to try to win back the money you lost? Never...1 Sometimes...2 Most of the time...3 Almost always...4 (DO NOT READ) Don't Know...8 (DO NOT READ) Refused / No Response...9 Q4. Thinking about the past 12 months, how often have you borrowed money or sold anything to get money to gamble? Never...1 Sometimes...2 Most of the time...3 Almost always...4 (DO NOT READ) Don't Know...8 (DO NOT READ) Refused / No Response...9

20 Page 20 Q5. Thinking about the past 12 months, how often have you felt that you might have a problem with gambling? Never...1 Sometimes...2 Most of the time...3 Almost always...4 (DO NOT READ) Don't Know...8 (DO NOT READ) Refused / No Response...9 Q6. Thinking about the past 12 months, how often have people criticized your betting or told you that you had a gambling problem, regardless of whether or not you thought it was true? Never...1 Sometimes...2 Most of the time...3 Almost always...4 (DO NOT READ) Don't Know...8 (DO NOT READ) Refused / No Response...9 Q7. Thinking about the past 12 months, how often have you felt guilty about the way you gamble, or what happens when you gamble? Never...1 Sometimes...2 Most of the time...3 Almost always...4 (DO NOT READ) Don't Know...8 (DO NOT READ) Refused / No Response...9 Q8. Thinking about the past 12 months, how often has your gambling caused you any health problems, including stress or anxiety? Never...1 Sometimes...2 Most of the time...3 Almost always...4 (DO NOT READ) Don't Know...8 (DO NOT READ) Refused / No Response...9