 |
 |
|
|
|
 Gambling is an increasingly accessible activity in most modern jurisdictions. Among the recent legalized gambling activities made available in the province of Quebec, it is worth mentioning the opening of a new casino in 1996 and the record number of video lottery terminals (n = 15 314) accessible in fiscal year 1998–1999 (1), compared with the terminals (n = 14 644) available in 1995–1996 (2). Despite the growing availability of legalized gambling, many Canadian provinces have not reported an increase in the prevalence of problem gambling. For example, the results from a 2001 study of nearly 5000 Ontario adults classified 0.7% of the population as PPGs (3), whereas the reported prevalence in 1995 was 2.0% (4). Manitoba also reported a similar trend, with a PPG prevalence of 1.9% in 1995 (5), compared with 1.1% in 2001 (6). In New Brunswick, the prevalence of PPGs was 2.2% in 1996 (7) and 1.4% in 2001 (8). Alberta reported that the combined proportion of ARGs and PPGs was 5.4 % in 1993 (9) and 4.8% in 1998 (10). Although these provincial studies were not conducted with the same procedures or by the same research teams, the data show that none of these provinces reported an increasing trend in gambling prevalence. Our study, which involved 2 phases, used a cross-sectional design to evaluate changes in the prevalence of pathological gambling in Quebec from 1996 to 2002. Phase 1 was designed to estimate the 2002 prevalence of ARGs and PPGs in the adult population of Quebec, to highlight certain characteristics associated with the gambling habits of adults in Quebec, to compare the 2002 prevalence with that estimated for 1996, and to compare the prevalence as captured by 2 screening instruments, the SOGS and the CPGI. Phase 2 aimed to compare the classifications obtained from the 2 instruments with classifications obtained through a clinical evaluation based on a semistructured telephone interview among the same participants. MethodPhase 1 Sample. Telephone numbers were drawn from a database of randomly generated phone numbers covering all regions of Quebec. Within each household, the selected participant was randomly chosen from among all persons aged 18 years and over. The employees of a survey firm carried out the telephone interviews between May and November 2002, and when necessary, 20 attempts were made to reach a number. A total of 8842 interviews were completed, with a response rate of 60.8% (Note 1). Participant responses were weighted according to the number of telephone call attempts to reach the resident, the number of adults living in each residence, the sex of the participant, the overall response rate, and the region (based on 17 sociomedical regions of Quebec and according to data from the 2001 census of Statistics Canada, 11). The average length of an interview was 11 minutes. Instruments. The telephone interview used in this study had 4 sections: 1) participation in gambling, 2) evaluation of problem gambling, 3) associated problems, and 4) general information. 1. Participation in gambling: This inventory determined the frequency of participation in various gambling activities, as well as the amount of money spent on these games. It was largely inspired by the inventory used in the CPGI (12). 2. Evaluation of problem gambling: The gambling activities and prevalence were assessed according to 2 screening instruments, the SOGS and the CPGI. The interviewers used the version of the SOGS (13) that is adapted for telephone interviews (14). The SOGS includes 20 items based on the DSM-III-R diagnostic criteria for pathological gambling. Participants who scored 0 to 2 were classified as NPGs; those who scored 3 or 4, ARGs; and those who scored 5 or more, PPGs. Since most participants were French-speaking, a translated version of the SOGS was used (15). The CPGI (12) has 14 items, 9 of which were used to calculate the overall index. Participants who scored between 0 and 3 were classified as NPGs, those who scored 4 to 7 were classified as ARGs, and those who scored 8 or more were classified as PPGs. The CPGI was developed in both French and English. The 2 instruments aim to estimate both lifetime and current-year (last 12 months) prevalence. Only results relating to current prevalence are reported here. 3. Associated problems: Eighteen questions used at the time of CPGI validation (12) evaluated gambling-related problems, drug addiction, and psychological distress. 4. General information: 22 questions gathered general information such as sex, age, and level of education. Procedure. Respondents had to answer “yes” to one of the following criteria to be assessed for problem gambling: 1) have spent more than $520 annually on gambling or 2) have played too much, spent too much money, or spent too much time gambling. Sections 3 (associated problems) and 4 (general information) were administered to all participants. However, some items of section 3 were not administered to participants who reported that they had not played. Phase 2 In Phase 2, we compared the participant classifications obtained in Phase 1 with classifications obtained from a semistructured clinical telephone interview conducted by a psychologist. Sample. At the end of the Phase 1 interview, some participants (a randomly selected number of NPG participants and all the ARG and PPG participants) were solicited to answer further questions during a second phone interview. Of the 1605 persons solicited, 952 (59%) agreed to receive a second call. This collaboration rate did not vary across the 3 participant categories (c2 = 4.33, df 2, 1605; P > 0.05). Respondents who agreed to be contacted again were also compared with those who refused on the variables of sex, age, level of education, and income. They differed only in age distribution (c2 = 10.74, df 4, 1456; P < 0.05): those who agreed to participate in a follow-up were proportionately more numerous in the group aged 25 to 34 years than in the other age groups. Of the 952 participants who agreed to a follow-up, only 201 were registered on the Phase 2 interview list: 120 NPGs (randomly selected from among 851 consenting respondents), 53 ARGs, and 28 PPGs. For reasons beyond our control, the survey firm did not transfer some files to the researchers, which caused the withdrawal of 3 ARG files from the 56 available and 17 PPG files from the 45 available on the list of participants who had accepted to be contacted a second time. Of the 201 participants who were called back, 169 were reached and 133 completed a telephone interview, resulting in a participation rate of 66% among those who were called back and an overall response rate of 39.3% for the second phase (Note 2). Persons who had refused to participate or who could not be reached did not differ from those who completed the clinical interview in relation to participant categories (NPG, ARG, and PPG), sex, education, or income. However, there was a difference in the age distribution (c2 = 9.55, df 4, 190; P < 0.05): those who did not complete the clinical interview (that is, who refused or were unreachable) were significantly more likely to be in the category aged 35 to 44 years than were those who completed the interview (P < 0.05). Procedure. Clinical psychologists trained to conduct the semistructured clinical telephone interviews evaluated gambling habits. Participants were contacted at a maximum interval of 3 months following the end of Phase 1. The psychologists made a clinical assessment based on the diagnostic criteria of pathological gambling according to the DSM-IV. They were blinded to the gambling habits of the participants and the scores obtained in Phase 1. ResultsPhase 1 In 2002, 4 adults out of 5 (81%) living in Quebec reported having gambled during the previous year. This percentage is lower than the 90% (Note 3) estimate observed in 1996 by Ladouceur and colleagues (16) (c2 = 59.13, df 1, 10 099, P < 0.001). Table 1 presents the 6 most popular categories of gambling activities within the general population, as well as the percentage of people who reported having participated in these gambling activities in the year leading up to the survey.
2002 Prevalence of Pathological Gambling We found the current prevalence of PPGs and of ARGs in 2002 to be 0.8% (95%CI, 0.6 to1.0) and 0.9% (95%CI, 0.7 to 1.1), respectively. Thus the number of adult PPGs in Quebec in 2002 ranged between 35 000 and 56 000, while the number of adult ARGs ranged between 40 000 and 62 000 (Table 2).
Sociodemographic Characteristics The relative proportions of the 3 categories of participants varied according to sex (c2 = 19.95, df 2, 8829; P < 0.001), age (c2 = 22.22, df 8, 7966; P < 0.01), level of education (c2 = 12.08, df 2, 8698; P < 0.01), and socioeconomic status (c2 = 22.55, df 4, 8586; P < 0.01). There were more men in the ARG (65%) and PPG (69%) categories, as compared with the NPG category (49%) (P < 0.05). The proportion of ARGs between the ages of 18 and 24 years (26%) was statistically higher than the proportions in the other participant categories (NPG and PPG both 12%). In addition, among persons aged 55 years and over, we found more without a gambling problem (27%) or in the ARG category (22%) than we did in the PPG category (13%) (P < 0.05). Among persons who did not complete their grade school or high school education, there were more PPGs (68%) than NPGs (49%) or ARGs (58%). Persons who reported being below the poverty line were more frequently classified as PPGs (29%) than as ARGs (17%) or NPGs (12%). Comparison of Prevalence in 1996 and in 2002 In 1996, 97.6% of the Quebec population were classified as NPGs and 1.4% as ARGs, whereas 1.0% were estimated to be PPGs. When we compared those current prevalence rates with the estimates obtained in 2002 (98.3%, 0.9%, and 0.8%, respectively), we found no statistical difference (c2 = 4.05, df 2, 10 086; P > 0.05). Prevalence Revealed by the SOGS and by the CPGI By comparing the results obtained from the SOGS with those obtained from the CPGI, we observed respective 12-month prevalence rates of 0.9% and 0.7% for PPGs, 0.9% and 1.0% for ARGs, and 98.3% and 98.2% for NPGs (see Table 2). A chi-square analysis of the 2 independent samples, comparing the 2 instruments, showed no significant difference in the prevalence estimates (c2 = 1.19, df 2, 8230; P > 0.05). Phase 2 Concordance Among Classifications. A test of marginal homogeneity for the paired observations indicates a significant difference between the classifications obtained by the 2 measures (MH = 53.0, df 3, 137; P < 0.0001). The rate of agreement between measures reaches 72% (98/137) but radically decreases to 25% (13/51) when only the number of ARGs and PPGs defined according to the SOGS or the CGPI are examined. The results show that 82% of the PPGs identified by the SOGS or the CPGI were not classified as pathological gamblers during the clinical interview. Both instruments provided similar proportions of misclassifications, compared with the clinical interview (SOGS, 78%; CPGI, 88%). With regard to the ARGs identified by the SOGS or the CPGI, 71% of participants did not maintain their classification in the clinical interview (the performance of the SOGS was similar to the CPGI: 67% and 72%, respectively). DiscussionThe prevalence of pathological gambling in Quebec in 2002 did not differ significantly from that observed 6 years earlier. Moreover, the 2002 prevalence closely resembles the findings of a recent Ontario study, which estimated the proportion of PPGs to be 0.7% (3). The lack of increase in the prevalence of pathological gambling in Quebec from 1996 to 2002 is similar to the trend observed in many of the other prevalence studies recently conducted in different Canadian provinces (3,6,8,10). Given the increasing availability of gambling, we identified several factors to explain the stability of the pathological gambling results over a 6-year period. First, we noticed a significant reduction in the rate of participation in gambling between 1996 and 2002, from 90% to 81%. Second, the different initiatives implemented in Quebec to reduce the incidence of problem gambling have perhaps started to slow down the progression of excessive gambling habits. Third, in 2001, the Quebec government started to implement a provincial treatment program for pathological gamblers. Specialized training was offered to more than 300 clinicians in 2001 and 2002, and free specialized services are now available in every region of Quebec. The combination of these efforts has perhaps contributed to stabilizing the development of problem gambling behaviours. Conversely, the absence of differences between the scores obtained using the CPGI and those obtained using the SOGS can be explained by the fact that the 2 instruments have the same goal, which is to measure problem gambling. Indeed, 55% of the questions in the CPGI and 80% of those in the SOGS deal with reporting negative consequences (17). Moreover, the correlations between the CPGI and the SOGS items vary between 0.63 and 0.72, and 2 of the 9 CPGI questions come from the SOGS (12). The final result to be discussed here relates to the discrepancy observed between the findings obtained using the screening instruments and those from the clinical interview. Are the prevalence rates revealed by the SOGS or the CPGI overestimated? Could these instruments have good sensitivity but lack specificity? Alternatively, could it be that the clinical interview lacks sensitivity and produces an underestimate? A similar discrepancy was previously pointed out in at least 3 empirical studies on gambling issues that showed a significant difference between the results of evaluations using a telephone survey or a questionnaire and those using a clinical evaluation (18–20). This discrepancy is not specific to the assessment of gambling behaviours. Similar results were found with studies of behaviour related to alcohol (21), sexuality (22), and drug and tobacco use (23,24). It is therefore clear that the type of instruments and the modality of their administration have a major influence on outcomes when different behaviours are being assessed. How can we ensure that our evaluation of gambling activity is accurate in the case of at-risk and pathological gamblers? More importantly, what do these figures mean? Future research will need to clarify these observations. Follow-up studies should continue to assess the development of problem gambling in the province of Quebec as they do elsewhere, by using cross-sectional and longitudinal study designs based on standardized, valid measures of pathological gambling and related activities. Funding and SupportThe research received financial support from the Ministère de la santé et des services sociaux du Québec. Notes1. The reported response rate is calculated as it was in studies conducted in 1989 and 1996: the rate of response = the number of individuals who completed the telephone interview / (individuals who completed the telephone interview + standard refusal + refusal for health problems). 2. The overall response rate of 39.3% corresponds to the product of the collaboration rate (952/1605 = 59.3%) by the interview participation rate (133/201 = 66.2%). 3. In 1996, a gambling participation rate of 63% was obtained from responses to the question, “In the past year, how many times have you gambled or bet with money?” graded on a 4-point Likert scale presented at the beginning of the telephone interview. However, subsequent interview responses to questions related to participation in 10 gambling activities during the past year increased the gambling rate to 90%. This is partly explained by the fact that most participants (71%) did not initially consider the purchase of lottery tickets (all types) to be gambling. We have thus used the rate of 90% for comparison with our 2002 data. References1. Loto-Québec 1998–1999 annual report. Quebec (QC): Loto-Québec; 1999. 2. Loto-Québec 1995–1996 annual report. Quebec (QC): Loto-Québec; 1996. 3. Wiebe J, Single E, Falkowski-Ham A. Measuring gambling and problem gambling in Ontario. Toronto (ON): Canadian Centre on Substance Abuse and Responsible Gambling Council; 2001. 4. Ferris J, Stirpe T. Gambling in Ontario: a report from a general population survey on gambling-related problems and opinions. Ontario: Addiction Research Foundation; 1995. 5. Criterion Research Corporation. Problem gambling study: final report. Report prepared for the Manitoba Lotteries Foundation. Winnipeg (MB): Criterion Research Corporation; 1995. 6. Brown D, Patton D, Dhaliwal J, Pankratz C, Broszeit B. Gambling involvement and problem gambling in Manitoba. Winnipeg (MB): Addictions Foundation of Manitoba; 2002. 7. Baseline Market Research. Final report: prevalence study. Problem gambling: wave 2. Fredericton (NB): New Brunswick Department of Finance; 1996. 8. Focal Research Consultants Ltd. 2001 survey of gambling and problem gambling in New Brunswick. Fredericton (NB): New Brunswick Department of Health and Wellness; 2001. 9. Wynne H, Smith G, Volberg RA. Adult gambling and problem gambling in Alberta. Edmonton: Alberta Lotteries and Gaming; 1994. 10. Wynne Resources. Adult gambling and problem gambling in Alberta. Edmonton: Alberta Alcohol and Drug Abuse Commission; 1998. 11. Statistics Canada. 2001 Census. Available: http://www12.statcan.ca/francais/census01/products/standard/themes/Index.cfm. Accessed 2003 July 30. 12. Ferris J, Wynne H.The Canadian Problem Gambling Index: final report. Ottawa: Canadian Centre on Substance Abuse; 2001. 13. Lesieur H, Blume SB. The South Oaks Gambling Screen (SOGS): a new instrument for the identification of pathological gamblers. Am J Psychiatry 1987;144:1184–8. 14. Volberg RA, Steadman HJ. Refining prevalence estimates of pathological gambling. Am J Psychiatry 1988;145;502–5. 15. Ladouceur R. Prevalence estimates of pathological gamblers in Quebec. Can J Psychiatry 1991;36:732–4. 16. Ladouceur R, Jacques C, Ferland F, Giroux I. Prevalence of problem gambling: a replication study 7 years later. Can J Psychiatry 1999;44:802–4. 17. Ladouceur R. A different approach to problem gambling and its impact on responsible gambling. Presented at the 2003 Discovery Congress; 2003 Sept 21–24; Toronto (ON). 18. Ladouceur R, Bouchard C, Rhéaume N, Jacques C, Ferland F, Leblond J, and others. Is the SOGS an accurate measure of pathological gambling among children, adolescents and adults? J Gambl Stud 2000;16:1–24. 19. Volberg RA, Bank SM. A review of two measures of pathological gambling in the United States. Journal of Gambling Behavior 1990;6:153–63. 20. Walker MB, Dickerson MG. The prevalence of problem and pathological gambling: a critical analysis. J Gambl Stud 1996;12:233–49. 21. Bisson J, Nadeau L, Demers A. The validity of the CAGE scale to screen for heavy drinking and drinking problems in a general population survey. Addiction 1999;94:715–22. 22. Gribble JN, Miller H, Roger SM, Turner CH. Interview mode and measurement of sexual behaviours: methodological issues. J Sex Res 1999;36:16–24. 23. Aquilino WS. Telephone versus face-to-face interviewing for household drug use surveys. Int J Addict 1992;27:71–91. 24. Gribble JN, Miller H, Cooley PC, Catania JA, Pollack L, Turner CH. The impact of T-ACASI interviewing on reported drug use among men who have sex with men. Subst Use Misuse 2000:35:869–90. Author(s)Manuscript received June 2004, revised, and accepted October 2004. 1. Professor, École de psychologie, Université Laval, Quebec City, Quebec. 2. Researcher, École de psychologie, Université Laval, Quebec City, Quebec. 3. Sociologist, Institut national de santé publique du Québec, Quebec City, Quebec. 4. Statistician, Institut national de santé publique du Québec, Quebec City, Quebec. Address for correspondence: Dr R Ladouceur, École de psychologie, Université Laval, Quebec, QC G1K 7P4 e-mail: Robert.Ladouceur@psy.ulaval.ca
1 | 2
|
|||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||